diff --git a/.gitattributes b/.gitattributes index a6344aac8c09253b3b630fb776ae94478aa0275b..26237b53eeed9669adb3d7f7e8bc5a33d2a01ed4 100644 --- a/.gitattributes +++ b/.gitattributes @@ -33,3 +33,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text *.zip filter=lfs diff=lfs merge=lfs -text *.zst filter=lfs diff=lfs merge=lfs -text *tfevents* filter=lfs diff=lfs merge=lfs -text +checkpoint-13500/tokenizer.json filter=lfs diff=lfs merge=lfs -text +checkpoint-13796/tokenizer.json filter=lfs diff=lfs merge=lfs -text +tokenizer.json filter=lfs diff=lfs merge=lfs -text diff --git a/chat_template.jinja b/chat_template.jinja new file mode 100644 index 0000000000000000000000000000000000000000..ed10d0cf20ed185c105baf9a5234ab076f0b4803 --- /dev/null +++ b/chat_template.jinja @@ -0,0 +1,37 @@ +[gMASK]{% for item in messages %}{% if item['tools'] is defined %}<|system|> +你是一个名为 ChatGLM 的人工智能助手。你是基于智谱AI训练的语言模型 GLM-4 模型开发的，你的任务是针对用户的问题和要求提供适当的答复和支持。 + +# 可用工具{% set tools = item['tools'] %}{% for tool in tools %}{% if tool['type'] == 'function' %} + +## {{ tool['function']['name'] }} + +{{ tool['function'] | tojson(indent=4) }} +在调用上述函数时，请使用 Json 格式表示调用的参数。{% elif tool['type'] == 'python' %} + +## python + +当你向 `python` 发送包含 Python 代码的消息时，该代码将会在一个有状态的 Jupyter notebook 环境中执行。 +`python` 返回代码执行的输出，或在执行 60 秒后返回超时。 +`/mnt/data` 将会持久化存储你的文件。在此会话中，`python` 无法访问互联网。不要使用 `python` 进行任何网络请求或者在线 API 调用，这些在线内容的访问将不会成功。{% elif tool['type'] == 'simple_browser' %} + +## simple_browser + +你可以使用 `simple_browser` 工具。该工具支持以下函数： +`search(query: str, recency_days: int)`：使用搜索引擎进行查询并显示结果，可以使用 `recency_days` 参数控制搜索内容的时效性。 +`mclick(ids: list[int])`：获取一系列指定 id 的页面内容。每次调用时，须选择3-10个页面。选择多个角度的页面，同时尽可能选择可信任的信息来源。考虑到部分页面是无法加载的，你也可以多打开一些可能有用的页面而不用担心内容过多。 +`open_url(url: str)`：打开指定的 URL。 + +使用 `【{引用 id}†{引用文本}】` 来引用内容。 + +操作步骤：1. 使用 `search` 来获得信息列表; 2. 使用 `mclick` 来获取指定 ID 页面的内容; 3. 根据获得的内容进行回复。在回复中应当引用信息来源。 + 如果用户提供了 URL，也可以用 `open_url` 直接打开页面。 +如果初次搜索结果没有找到合适的信息，也可以再次使用 `search` 进行搜索。{% elif tool['type'] == 'cogview' %} + +## cogview + +如果用户的请求中包含了对图像的描述，你可以使用 `cogview` 来生成图像并展示给用户。你需要向 `cogview` 发送图像描述，规则： +- 发送给 `cogview` 的消息必须使用英语。用户的中文描述必须完全翻译为英语。 +- 应当尽可能详细地描述图像生成的需求，需求描述约 100 英文单词。 +- 保持用户原始描述的意图。不要虚构内容或者没见过的人物。 +- 如无特殊说明，所在地为中国，持有中国立场并遵循中国社会主义价值观。{% endif %}{% endfor %}{% endif %}{% if item['content'] %}<|{{ item['role'] }}|>{{ item['metadata'] }} +{{ item['content'] }}{% endif %}{% endfor %}{% if add_generation_prompt %}<|assistant|>{% endif %} \ No newline at end of file diff --git a/checkpoint-13500/chat_template.jinja b/checkpoint-13500/chat_template.jinja new file mode 100644 index 0000000000000000000000000000000000000000..ed10d0cf20ed185c105baf9a5234ab076f0b4803 --- /dev/null +++ b/checkpoint-13500/chat_template.jinja @@ -0,0 +1,37 @@ +[gMASK]{% for item in messages %}{% if item['tools'] is defined %}<|system|> +你是一个名为 ChatGLM 的人工智能助手。你是基于智谱AI训练的语言模型 GLM-4 模型开发的，你的任务是针对用户的问题和要求提供适当的答复和支持。 + +# 可用工具{% set tools = item['tools'] %}{% for tool in tools %}{% if tool['type'] == 'function' %} + +## {{ tool['function']['name'] }} + +{{ tool['function'] | tojson(indent=4) }} +在调用上述函数时，请使用 Json 格式表示调用的参数。{% elif tool['type'] == 'python' %} + +## python + +当你向 `python` 发送包含 Python 代码的消息时，该代码将会在一个有状态的 Jupyter notebook 环境中执行。 +`python` 返回代码执行的输出，或在执行 60 秒后返回超时。 +`/mnt/data` 将会持久化存储你的文件。在此会话中，`python` 无法访问互联网。不要使用 `python` 进行任何网络请求或者在线 API 调用，这些在线内容的访问将不会成功。{% elif tool['type'] == 'simple_browser' %} + +## simple_browser + +你可以使用 `simple_browser` 工具。该工具支持以下函数： +`search(query: str, recency_days: int)`：使用搜索引擎进行查询并显示结果，可以使用 `recency_days` 参数控制搜索内容的时效性。 +`mclick(ids: list[int])`：获取一系列指定 id 的页面内容。每次调用时，须选择3-10个页面。选择多个角度的页面，同时尽可能选择可信任的信息来源。考虑到部分页面是无法加载的，你也可以多打开一些可能有用的页面而不用担心内容过多。 +`open_url(url: str)`：打开指定的 URL。 + +使用 `【{引用 id}†{引用文本}】` 来引用内容。 + +操作步骤：1. 使用 `search` 来获得信息列表; 2. 使用 `mclick` 来获取指定 ID 页面的内容; 3. 根据获得的内容进行回复。在回复中应当引用信息来源。 + 如果用户提供了 URL，也可以用 `open_url` 直接打开页面。 +如果初次搜索结果没有找到合适的信息，也可以再次使用 `search` 进行搜索。{% elif tool['type'] == 'cogview' %} + +## cogview + +如果用户的请求中包含了对图像的描述，你可以使用 `cogview` 来生成图像并展示给用户。你需要向 `cogview` 发送图像描述，规则： +- 发送给 `cogview` 的消息必须使用英语。用户的中文描述必须完全翻译为英语。 +- 应当尽可能详细地描述图像生成的需求，需求描述约 100 英文单词。 +- 保持用户原始描述的意图。不要虚构内容或者没见过的人物。 +- 如无特殊说明，所在地为中国，持有中国立场并遵循中国社会主义价值观。{% endif %}{% endfor %}{% endif %}{% if item['content'] %}<|{{ item['role'] }}|>{{ item['metadata'] }} +{{ item['content'] }}{% endif %}{% endfor %}{% if add_generation_prompt %}<|assistant|>{% endif %} \ No newline at end of file diff --git a/checkpoint-13500/config.json b/checkpoint-13500/config.json new file mode 100644 index 0000000000000000000000000000000000000000..bc35575d2212a40a1ecaf26ea55c3bdf0649ddfb --- /dev/null +++ b/checkpoint-13500/config.json @@ -0,0 +1,31 @@ +{ + "architectures": [ + "GlmForCausalLM" + ], + "attention_bias": true, + "attention_dropout": 0.0, + "eos_token_id": [ + 151329, + 151336, + 151338 + ], + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 4096, + "initializer_range": 0.02, + "intermediate_size": 13696, + "max_position_embeddings": 131072, + "model_type": "glm", + "num_attention_heads": 32, + "num_hidden_layers": 40, + "num_key_value_heads": 2, + "pad_token_id": 151329, + "partial_rotary_factor": 0.5, + "rms_norm_eps": 1.5625e-07, + "rope_theta": 10000.0, + "tie_word_embeddings": false, + "torch_dtype": "bfloat16", + "transformers_version": "4.53.1", + "use_cache": true, + "vocab_size": 151552 +} diff --git a/checkpoint-13500/generation_config.json b/checkpoint-13500/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..d5a9c67e4e2fe314feb0a9b23144f0efeb644787 --- /dev/null +++ b/checkpoint-13500/generation_config.json @@ -0,0 +1,10 @@ +{ + "_from_model_config": true, + "eos_token_id": [ + 151329, + 151336, + 151338 + ], + "pad_token_id": 151329, + "transformers_version": "4.53.1" +} diff --git a/checkpoint-13500/model-00001-of-00004.safetensors b/checkpoint-13500/model-00001-of-00004.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9bea7fe6d9e5621bd235db1c2f71cb287126f37a --- /dev/null +++ b/checkpoint-13500/model-00001-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:155826036741998a0994aa41cc004b8d033d258d0e1eb9b4d353d98793e0ade1 +size 4984133600 diff --git a/checkpoint-13500/model-00002-of-00004.safetensors b/checkpoint-13500/model-00002-of-00004.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..6bc3bb7ef9505549aa33a1f654010312c00e13a8 --- /dev/null +++ b/checkpoint-13500/model-00002-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:da45d4f3da401b2fd1e3799b29c0d53e4b191dc42d4522ab1ce210779aa18005 +size 4895075168 diff --git a/checkpoint-13500/model-00003-of-00004.safetensors b/checkpoint-13500/model-00003-of-00004.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0681278019f935a4b63927fff19d7588a18afdaa --- /dev/null +++ b/checkpoint-13500/model-00003-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a4e925c3c4f4a6e0a7d06e04dde933a72b0317cc752b320f6decd02e56189c1f +size 4895075184 diff --git a/checkpoint-13500/model-00004-of-00004.safetensors b/checkpoint-13500/model-00004-of-00004.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b3c66a64da9f4687f7701291b3d97987606eb5ae --- /dev/null +++ b/checkpoint-13500/model-00004-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a328b41622b5658ba9e372df9f73a422e363adcc9adfe6344ac8f64f3d042361 +size 4025669744 diff --git a/checkpoint-13500/model.safetensors.index.json b/checkpoint-13500/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..63641162b53de18c620d9bdad367eb87482d7fbf --- /dev/null +++ b/checkpoint-13500/model.safetensors.index.json @@ -0,0 +1,451 @@ +{ + "metadata": { + "total_parameters": 9399951360, + "total_size": 18799902720 + }, + "weight_map": { + "lm_head.weight": "model-00004-of-00004.safetensors", + "model.embed_tokens.weight": "model-00001-of-00004.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.30.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.31.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.32.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.33.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.33.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.33.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.33.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.34.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.34.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.34.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.34.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.35.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.36.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.37.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.38.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.39.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.norm.weight": "model-00004-of-00004.safetensors" + } +} diff --git a/checkpoint-13500/optimizer.pt b/checkpoint-13500/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..4f798dfc5da1874cc455fcb4256ead99b6512138 --- /dev/null +++ b/checkpoint-13500/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f80025c2601541bcd3889bea1fbb5d3ec4780e007d90b25c7dc9d85611f11da9 +size 37600193476 diff --git a/checkpoint-13500/rng_state.pth b/checkpoint-13500/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..1feba1a6538e93b94696d3773853dbc8947b0cad --- /dev/null +++ b/checkpoint-13500/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:098b29492211804ab324a36f37466821d948280bb74fce4ba895c03f13ecd878 +size 14645 diff --git a/checkpoint-13500/scheduler.pt b/checkpoint-13500/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..e8b7eb372ec280b2f7169acbe51d9be519972546 --- /dev/null +++ b/checkpoint-13500/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0735265f475bf172092ac7bde32480273b75b9874336b2e6204a21df57b8680b +size 1465 diff --git a/checkpoint-13500/special_tokens_map.json b/checkpoint-13500/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..582da4ad333feaaa0c427b4678a3e48447ea932f --- /dev/null +++ b/checkpoint-13500/special_tokens_map.json @@ -0,0 +1,32 @@ +{ + "additional_special_tokens": [ + "<|endoftext|>", + "[MASK]", + "[gMASK]", + "[sMASK]", + "", + "", + "<|system|>", + "<|user|>", + "<|assistant|>", + "<|observation|>", + "<|begin_of_image|>", + "<|end_of_image|>", + "<|begin_of_video|>", + "<|end_of_video|>" + ], + "eos_token": { + "content": "<|endoftext|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "<|endoftext|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoint-13500/tokenizer.json b/checkpoint-13500/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..2c24a7b2180f4b088af2671a705217ba4302fee5 --- /dev/null +++ b/checkpoint-13500/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8a7269d6daa6328de533def0082ff9d3a825bb89036cbcc665c324f941f67fbf +size 19967863 diff --git a/checkpoint-13500/tokenizer_config.json b/checkpoint-13500/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..38214260739a871e4c13601f88d904a2ad1fd601 --- /dev/null +++ b/checkpoint-13500/tokenizer_config.json @@ -0,0 +1,145 @@ +{ + "added_tokens_decoder": { + "151329": { + "content": "<|endoftext|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151330": { + "content": "[MASK]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151331": { + "content": "[gMASK]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151332": { + "content": "[sMASK]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151333": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151334": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151335": { + "content": "<|system|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151336": { + "content": "<|user|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151337": { + "content": "<|assistant|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151338": { + "content": "<|observation|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151339": { + "content": "<|begin_of_image|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151340": { + "content": "<|end_of_image|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151341": { + "content": "<|begin_of_video|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151342": { + "content": "<|end_of_video|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "additional_special_tokens": [ + "<|endoftext|>", + "[MASK]", + "[gMASK]", + "[sMASK]", + "", + "", + "<|system|>", + "<|user|>", + "<|assistant|>", + "<|observation|>", + "<|begin_of_image|>", + "<|end_of_image|>", + "<|begin_of_video|>", + "<|end_of_video|>" + ], + "clean_up_tokenization_spaces": false, + "do_lower_case": false, + "eos_token": "<|endoftext|>", + "extra_special_tokens": {}, + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 128000, + "pad_token": "<|endoftext|>", + "padding_side": "left", + "remove_space": false, + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoint-13500/trainer_state.json b/checkpoint-13500/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..28ac8f8f09d34ddf765a2e01f80c4e5307071c53 --- /dev/null +++ b/checkpoint-13500/trainer_state.json @@ -0,0 +1,9484 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 1.957123699757167, + "eval_steps": 500, + "global_step": 13500, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.0014497481062665362, + "grad_norm": 31.75, + "learning_rate": 1.9986952739924617e-05, + "loss": 1.9642, + "step": 10 + }, + { + "epoch": 0.0028994962125330724, + "grad_norm": 10.5, + "learning_rate": 1.99724557842853e-05, + "loss": 2.0457, + "step": 20 + }, + { + "epoch": 0.004349244318799609, + "grad_norm": 8.0, + "learning_rate": 1.9957958828645986e-05, + "loss": 1.7015, + "step": 30 + }, + { + "epoch": 0.005798992425066145, + "grad_norm": 10.4375, + "learning_rate": 1.994346187300667e-05, + "loss": 1.6103, + "step": 40 + }, + { + "epoch": 0.007248740531332681, + "grad_norm": 9.5, + "learning_rate": 1.9928964917367355e-05, + "loss": 1.7592, + "step": 50 + }, + { + "epoch": 0.008698488637599217, + "grad_norm": 5.09375, + "learning_rate": 1.991446796172804e-05, + "loss": 1.114, + "step": 60 + }, + { + "epoch": 0.010148236743865753, + "grad_norm": 5.625, + "learning_rate": 1.9899971006088723e-05, + "loss": 1.1614, + "step": 70 + }, + { + "epoch": 0.01159798485013229, + "grad_norm": 32.5, + "learning_rate": 1.9885474050449408e-05, + "loss": 1.1853, + "step": 80 + }, + { + "epoch": 0.013047732956398826, + "grad_norm": 4.5, + "learning_rate": 1.9870977094810092e-05, + "loss": 1.1573, + "step": 90 + }, + { + "epoch": 0.014497481062665362, + "grad_norm": 6.375, + "learning_rate": 1.9856480139170777e-05, + "loss": 1.183, + "step": 100 + }, + { + "epoch": 0.015947229168931897, + "grad_norm": 4.46875, + "learning_rate": 1.9841983183531458e-05, + "loss": 1.1313, + "step": 110 + }, + { + "epoch": 0.017396977275198434, + "grad_norm": 7.15625, + "learning_rate": 1.9827486227892142e-05, + "loss": 1.0763, + "step": 120 + }, + { + "epoch": 0.01884672538146497, + "grad_norm": 7.0625, + "learning_rate": 1.981298927225283e-05, + "loss": 1.0995, + "step": 130 + }, + { + "epoch": 0.020296473487731507, + "grad_norm": 4.03125, + "learning_rate": 1.9798492316613514e-05, + "loss": 1.0671, + "step": 140 + }, + { + "epoch": 0.02174622159399804, + "grad_norm": 4.5, + "learning_rate": 1.97839953609742e-05, + "loss": 1.0405, + "step": 150 + }, + { + "epoch": 0.02319596970026458, + "grad_norm": 4.15625, + "learning_rate": 1.9769498405334883e-05, + "loss": 1.0388, + "step": 160 + }, + { + "epoch": 0.024645717806531114, + "grad_norm": 4.28125, + "learning_rate": 1.9755001449695564e-05, + "loss": 1.025, + "step": 170 + }, + { + "epoch": 0.02609546591279765, + "grad_norm": 4.25, + "learning_rate": 1.974050449405625e-05, + "loss": 1.0053, + "step": 180 + }, + { + "epoch": 0.027545214019064186, + "grad_norm": 4.15625, + "learning_rate": 1.9726007538416933e-05, + "loss": 0.9725, + "step": 190 + }, + { + "epoch": 0.028994962125330724, + "grad_norm": 3.96875, + "learning_rate": 1.9711510582777617e-05, + "loss": 0.9925, + "step": 200 + }, + { + "epoch": 0.03044471023159726, + "grad_norm": 3.84375, + "learning_rate": 1.9697013627138302e-05, + "loss": 0.9483, + "step": 210 + }, + { + "epoch": 0.03189445833786379, + "grad_norm": 5.71875, + "learning_rate": 1.9682516671498986e-05, + "loss": 0.9845, + "step": 220 + }, + { + "epoch": 0.03334420644413033, + "grad_norm": 302.0, + "learning_rate": 1.966801971585967e-05, + "loss": 5.6501, + "step": 230 + }, + { + "epoch": 0.03479395455039687, + "grad_norm": 40.25, + "learning_rate": 1.9653522760220355e-05, + "loss": 7.2091, + "step": 240 + }, + { + "epoch": 0.03624370265666341, + "grad_norm": 12.625, + "learning_rate": 1.963902580458104e-05, + "loss": 6.3659, + "step": 250 + }, + { + "epoch": 0.03769345076292994, + "grad_norm": 27.625, + "learning_rate": 1.9624528848941724e-05, + "loss": 6.4682, + "step": 260 + }, + { + "epoch": 0.039143198869196476, + "grad_norm": 14.4375, + "learning_rate": 1.961003189330241e-05, + "loss": 6.3851, + "step": 270 + }, + { + "epoch": 0.040592946975463014, + "grad_norm": 12.5625, + "learning_rate": 1.9595534937663093e-05, + "loss": 6.0976, + "step": 280 + }, + { + "epoch": 0.04204269508172955, + "grad_norm": 15.1875, + "learning_rate": 1.9581037982023777e-05, + "loss": 6.1159, + "step": 290 + }, + { + "epoch": 0.04349244318799608, + "grad_norm": 14.5625, + "learning_rate": 1.956654102638446e-05, + "loss": 6.0918, + "step": 300 + }, + { + "epoch": 0.04494219129426262, + "grad_norm": 140.0, + "learning_rate": 1.9552044070745143e-05, + "loss": 6.0004, + "step": 310 + }, + { + "epoch": 0.04639193940052916, + "grad_norm": 79.0, + "learning_rate": 1.953754711510583e-05, + "loss": 6.0565, + "step": 320 + }, + { + "epoch": 0.047841687506795696, + "grad_norm": 65.0, + "learning_rate": 1.9523050159466515e-05, + "loss": 5.7795, + "step": 330 + }, + { + "epoch": 0.04929143561306223, + "grad_norm": 161.0, + "learning_rate": 1.95085532038272e-05, + "loss": 5.4378, + "step": 340 + }, + { + "epoch": 0.050741183719328765, + "grad_norm": 132.0, + "learning_rate": 1.9494056248187884e-05, + "loss": 5.024, + "step": 350 + }, + { + "epoch": 0.0521909318255953, + "grad_norm": 118.5, + "learning_rate": 1.9479559292548568e-05, + "loss": 4.4482, + "step": 360 + }, + { + "epoch": 0.05364067993186184, + "grad_norm": 107.5, + "learning_rate": 1.946506233690925e-05, + "loss": 4.1242, + "step": 370 + }, + { + "epoch": 0.05509042803812837, + "grad_norm": 129.0, + "learning_rate": 1.9450565381269934e-05, + "loss": 3.7731, + "step": 380 + }, + { + "epoch": 0.05654017614439491, + "grad_norm": 117.5, + "learning_rate": 1.9436068425630618e-05, + "loss": 3.6193, + "step": 390 + }, + { + "epoch": 0.05798992425066145, + "grad_norm": 206.0, + "learning_rate": 1.9421571469991302e-05, + "loss": 3.401, + "step": 400 + }, + { + "epoch": 0.059439672356927986, + "grad_norm": 185.0, + "learning_rate": 1.9407074514351987e-05, + "loss": 3.3099, + "step": 410 + }, + { + "epoch": 0.06088942046319452, + "grad_norm": 106.5, + "learning_rate": 1.9392577558712675e-05, + "loss": 3.2412, + "step": 420 + }, + { + "epoch": 0.062339168569461055, + "grad_norm": 163.0, + "learning_rate": 1.9378080603073356e-05, + "loss": 2.9854, + "step": 430 + }, + { + "epoch": 0.06378891667572759, + "grad_norm": 136.0, + "learning_rate": 1.936358364743404e-05, + "loss": 2.9673, + "step": 440 + }, + { + "epoch": 0.06523866478199412, + "grad_norm": 197.0, + "learning_rate": 1.9349086691794724e-05, + "loss": 3.0343, + "step": 450 + }, + { + "epoch": 0.06668841288826066, + "grad_norm": 189.0, + "learning_rate": 1.933458973615541e-05, + "loss": 2.9598, + "step": 460 + }, + { + "epoch": 0.0681381609945272, + "grad_norm": 177.0, + "learning_rate": 1.9320092780516093e-05, + "loss": 2.8282, + "step": 470 + }, + { + "epoch": 0.06958790910079374, + "grad_norm": 108.0, + "learning_rate": 1.9305595824876778e-05, + "loss": 2.8748, + "step": 480 + }, + { + "epoch": 0.07103765720706028, + "grad_norm": 139.0, + "learning_rate": 1.9291098869237462e-05, + "loss": 2.7971, + "step": 490 + }, + { + "epoch": 0.07248740531332681, + "grad_norm": 99.0, + "learning_rate": 1.9276601913598147e-05, + "loss": 2.7033, + "step": 500 + }, + { + "epoch": 0.07393715341959335, + "grad_norm": 187.0, + "learning_rate": 1.9262104957958828e-05, + "loss": 2.747, + "step": 510 + }, + { + "epoch": 0.07538690152585988, + "grad_norm": 114.0, + "learning_rate": 1.9247608002319515e-05, + "loss": 2.6358, + "step": 520 + }, + { + "epoch": 0.07683664963212641, + "grad_norm": 123.5, + "learning_rate": 1.92331110466802e-05, + "loss": 2.6722, + "step": 530 + }, + { + "epoch": 0.07828639773839295, + "grad_norm": 225.0, + "learning_rate": 1.9218614091040884e-05, + "loss": 2.5509, + "step": 540 + }, + { + "epoch": 0.07973614584465949, + "grad_norm": 106.0, + "learning_rate": 1.920411713540157e-05, + "loss": 2.5773, + "step": 550 + }, + { + "epoch": 0.08118589395092603, + "grad_norm": 115.5, + "learning_rate": 1.9189620179762253e-05, + "loss": 2.4187, + "step": 560 + }, + { + "epoch": 0.08263564205719257, + "grad_norm": 129.0, + "learning_rate": 1.9175123224122934e-05, + "loss": 2.4331, + "step": 570 + }, + { + "epoch": 0.0840853901634591, + "grad_norm": 115.5, + "learning_rate": 1.916062626848362e-05, + "loss": 2.388, + "step": 580 + }, + { + "epoch": 0.08553513826972564, + "grad_norm": 141.0, + "learning_rate": 1.9146129312844303e-05, + "loss": 2.4459, + "step": 590 + }, + { + "epoch": 0.08698488637599217, + "grad_norm": 162.0, + "learning_rate": 1.9131632357204987e-05, + "loss": 2.4265, + "step": 600 + }, + { + "epoch": 0.0884346344822587, + "grad_norm": 168.0, + "learning_rate": 1.9117135401565672e-05, + "loss": 2.367, + "step": 610 + }, + { + "epoch": 0.08988438258852524, + "grad_norm": 127.0, + "learning_rate": 1.910263844592636e-05, + "loss": 2.2841, + "step": 620 + }, + { + "epoch": 0.09133413069479178, + "grad_norm": 127.0, + "learning_rate": 1.908814149028704e-05, + "loss": 2.5091, + "step": 630 + }, + { + "epoch": 0.09278387880105832, + "grad_norm": 126.5, + "learning_rate": 1.9073644534647725e-05, + "loss": 2.2486, + "step": 640 + }, + { + "epoch": 0.09423362690732486, + "grad_norm": 100.5, + "learning_rate": 1.905914757900841e-05, + "loss": 2.3366, + "step": 650 + }, + { + "epoch": 0.09568337501359139, + "grad_norm": 135.0, + "learning_rate": 1.9044650623369094e-05, + "loss": 2.3301, + "step": 660 + }, + { + "epoch": 0.09713312311985793, + "grad_norm": 136.0, + "learning_rate": 1.9030153667729778e-05, + "loss": 2.3708, + "step": 670 + }, + { + "epoch": 0.09858287122612445, + "grad_norm": 118.0, + "learning_rate": 1.9015656712090463e-05, + "loss": 2.2925, + "step": 680 + }, + { + "epoch": 0.10003261933239099, + "grad_norm": 131.0, + "learning_rate": 1.9001159756451147e-05, + "loss": 2.2805, + "step": 690 + }, + { + "epoch": 0.10148236743865753, + "grad_norm": 93.0, + "learning_rate": 1.898666280081183e-05, + "loss": 2.2899, + "step": 700 + }, + { + "epoch": 0.10293211554492407, + "grad_norm": 135.0, + "learning_rate": 1.8972165845172516e-05, + "loss": 2.0617, + "step": 710 + }, + { + "epoch": 0.1043818636511906, + "grad_norm": 136.0, + "learning_rate": 1.89576688895332e-05, + "loss": 2.2515, + "step": 720 + }, + { + "epoch": 0.10583161175745714, + "grad_norm": 178.0, + "learning_rate": 1.8943171933893885e-05, + "loss": 2.1664, + "step": 730 + }, + { + "epoch": 0.10728135986372368, + "grad_norm": 158.0, + "learning_rate": 1.892867497825457e-05, + "loss": 2.1863, + "step": 740 + }, + { + "epoch": 0.10873110796999022, + "grad_norm": 133.0, + "learning_rate": 1.8914178022615254e-05, + "loss": 2.169, + "step": 750 + }, + { + "epoch": 0.11018085607625674, + "grad_norm": 111.5, + "learning_rate": 1.8899681066975938e-05, + "loss": 2.1845, + "step": 760 + }, + { + "epoch": 0.11163060418252328, + "grad_norm": 101.5, + "learning_rate": 1.8885184111336622e-05, + "loss": 2.1846, + "step": 770 + }, + { + "epoch": 0.11308035228878982, + "grad_norm": 112.0, + "learning_rate": 1.8870687155697303e-05, + "loss": 2.1391, + "step": 780 + }, + { + "epoch": 0.11453010039505636, + "grad_norm": 135.0, + "learning_rate": 1.8856190200057988e-05, + "loss": 2.1059, + "step": 790 + }, + { + "epoch": 0.1159798485013229, + "grad_norm": 256.0, + "learning_rate": 1.8841693244418672e-05, + "loss": 2.2222, + "step": 800 + }, + { + "epoch": 0.11742959660758943, + "grad_norm": 174.0, + "learning_rate": 1.8827196288779357e-05, + "loss": 2.0608, + "step": 810 + }, + { + "epoch": 0.11887934471385597, + "grad_norm": 140.0, + "learning_rate": 1.8812699333140045e-05, + "loss": 2.0371, + "step": 820 + }, + { + "epoch": 0.12032909282012251, + "grad_norm": 159.0, + "learning_rate": 1.8798202377500726e-05, + "loss": 2.045, + "step": 830 + }, + { + "epoch": 0.12177884092638903, + "grad_norm": 103.5, + "learning_rate": 1.878370542186141e-05, + "loss": 1.9997, + "step": 840 + }, + { + "epoch": 0.12322858903265557, + "grad_norm": 121.0, + "learning_rate": 1.8769208466222094e-05, + "loss": 2.0391, + "step": 850 + }, + { + "epoch": 0.12467833713892211, + "grad_norm": 122.5, + "learning_rate": 1.875471151058278e-05, + "loss": 2.062, + "step": 860 + }, + { + "epoch": 0.12612808524518865, + "grad_norm": 164.0, + "learning_rate": 1.8740214554943463e-05, + "loss": 2.1222, + "step": 870 + }, + { + "epoch": 0.12757783335145517, + "grad_norm": 110.5, + "learning_rate": 1.8725717599304148e-05, + "loss": 2.0443, + "step": 880 + }, + { + "epoch": 0.12902758145772172, + "grad_norm": 97.0, + "learning_rate": 1.8711220643664832e-05, + "loss": 1.9237, + "step": 890 + }, + { + "epoch": 0.13047732956398825, + "grad_norm": 103.5, + "learning_rate": 1.8696723688025516e-05, + "loss": 1.9752, + "step": 900 + }, + { + "epoch": 0.1319270776702548, + "grad_norm": 109.5, + "learning_rate": 1.86822267323862e-05, + "loss": 2.0141, + "step": 910 + }, + { + "epoch": 0.13337682577652132, + "grad_norm": 132.0, + "learning_rate": 1.8667729776746885e-05, + "loss": 1.9615, + "step": 920 + }, + { + "epoch": 0.13482657388278788, + "grad_norm": 102.0, + "learning_rate": 1.865323282110757e-05, + "loss": 1.9503, + "step": 930 + }, + { + "epoch": 0.1362763219890544, + "grad_norm": 117.5, + "learning_rate": 1.8638735865468254e-05, + "loss": 1.8662, + "step": 940 + }, + { + "epoch": 0.13772607009532095, + "grad_norm": 133.0, + "learning_rate": 1.862423890982894e-05, + "loss": 1.965, + "step": 950 + }, + { + "epoch": 0.13917581820158748, + "grad_norm": 114.0, + "learning_rate": 1.8609741954189623e-05, + "loss": 1.8814, + "step": 960 + }, + { + "epoch": 0.140625566307854, + "grad_norm": 140.0, + "learning_rate": 1.8595244998550307e-05, + "loss": 1.9293, + "step": 970 + }, + { + "epoch": 0.14207531441412055, + "grad_norm": 121.5, + "learning_rate": 1.858074804291099e-05, + "loss": 1.9772, + "step": 980 + }, + { + "epoch": 0.14352506252038708, + "grad_norm": 120.0, + "learning_rate": 1.8566251087271673e-05, + "loss": 1.9769, + "step": 990 + }, + { + "epoch": 0.14497481062665363, + "grad_norm": 70.5, + "learning_rate": 1.8551754131632357e-05, + "loss": 1.9106, + "step": 1000 + }, + { + "epoch": 0.14642455873292015, + "grad_norm": 132.0, + "learning_rate": 1.853725717599304e-05, + "loss": 1.9397, + "step": 1010 + }, + { + "epoch": 0.1478743068391867, + "grad_norm": 120.5, + "learning_rate": 1.852276022035373e-05, + "loss": 1.8997, + "step": 1020 + }, + { + "epoch": 0.14932405494545323, + "grad_norm": 110.5, + "learning_rate": 1.8508263264714414e-05, + "loss": 1.8802, + "step": 1030 + }, + { + "epoch": 0.15077380305171975, + "grad_norm": 130.0, + "learning_rate": 1.8493766309075095e-05, + "loss": 1.9104, + "step": 1040 + }, + { + "epoch": 0.1522235511579863, + "grad_norm": 127.0, + "learning_rate": 1.847926935343578e-05, + "loss": 1.8504, + "step": 1050 + }, + { + "epoch": 0.15367329926425283, + "grad_norm": 103.0, + "learning_rate": 1.8464772397796464e-05, + "loss": 1.8127, + "step": 1060 + }, + { + "epoch": 0.15512304737051938, + "grad_norm": 113.0, + "learning_rate": 1.8450275442157148e-05, + "loss": 1.8785, + "step": 1070 + }, + { + "epoch": 0.1565727954767859, + "grad_norm": 98.5, + "learning_rate": 1.8435778486517833e-05, + "loss": 1.7915, + "step": 1080 + }, + { + "epoch": 0.15802254358305245, + "grad_norm": 428.0, + "learning_rate": 1.8421281530878517e-05, + "loss": 3.0407, + "step": 1090 + }, + { + "epoch": 0.15947229168931898, + "grad_norm": 362.0, + "learning_rate": 1.84067845752392e-05, + "loss": 2.29, + "step": 1100 + }, + { + "epoch": 0.16092203979558553, + "grad_norm": 1312.0, + "learning_rate": 1.8392287619599886e-05, + "loss": 2.7752, + "step": 1110 + }, + { + "epoch": 0.16237178790185205, + "grad_norm": 576.0, + "learning_rate": 1.837779066396057e-05, + "loss": 2.8033, + "step": 1120 + }, + { + "epoch": 0.16382153600811858, + "grad_norm": 153.0, + "learning_rate": 1.8363293708321255e-05, + "loss": 2.5284, + "step": 1130 + }, + { + "epoch": 0.16527128411438513, + "grad_norm": 454.0, + "learning_rate": 1.834879675268194e-05, + "loss": 3.7134, + "step": 1140 + }, + { + "epoch": 0.16672103222065165, + "grad_norm": 2008.0, + "learning_rate": 1.8334299797042623e-05, + "loss": 2.2115, + "step": 1150 + }, + { + "epoch": 0.1681707803269182, + "grad_norm": 298.0, + "learning_rate": 1.8319802841403308e-05, + "loss": 1.7636, + "step": 1160 + }, + { + "epoch": 0.16962052843318473, + "grad_norm": 6.65625, + "learning_rate": 1.8305305885763992e-05, + "loss": 1.2321, + "step": 1170 + }, + { + "epoch": 0.17107027653945128, + "grad_norm": 5.15625, + "learning_rate": 1.8290808930124673e-05, + "loss": 1.0327, + "step": 1180 + }, + { + "epoch": 0.1725200246457178, + "grad_norm": 9.5, + "learning_rate": 1.8276311974485358e-05, + "loss": 1.0053, + "step": 1190 + }, + { + "epoch": 0.17396977275198433, + "grad_norm": 10.125, + "learning_rate": 1.8261815018846042e-05, + "loss": 1.0167, + "step": 1200 + }, + { + "epoch": 0.17541952085825088, + "grad_norm": 4.1875, + "learning_rate": 1.8247318063206727e-05, + "loss": 1.0274, + "step": 1210 + }, + { + "epoch": 0.1768692689645174, + "grad_norm": 4.625, + "learning_rate": 1.8232821107567414e-05, + "loss": 1.0167, + "step": 1220 + }, + { + "epoch": 0.17831901707078396, + "grad_norm": 3.984375, + "learning_rate": 1.82183241519281e-05, + "loss": 1.0135, + "step": 1230 + }, + { + "epoch": 0.17976876517705048, + "grad_norm": 13.1875, + "learning_rate": 1.820382719628878e-05, + "loss": 1.0732, + "step": 1240 + }, + { + "epoch": 0.18121851328331703, + "grad_norm": 4.28125, + "learning_rate": 1.8189330240649464e-05, + "loss": 1.0478, + "step": 1250 + }, + { + "epoch": 0.18266826138958356, + "grad_norm": 4.375, + "learning_rate": 1.817483328501015e-05, + "loss": 0.9829, + "step": 1260 + }, + { + "epoch": 0.18411800949585008, + "grad_norm": 3.96875, + "learning_rate": 1.8160336329370833e-05, + "loss": 1.0159, + "step": 1270 + }, + { + "epoch": 0.18556775760211663, + "grad_norm": 3.859375, + "learning_rate": 1.8145839373731518e-05, + "loss": 1.0163, + "step": 1280 + }, + { + "epoch": 0.18701750570838316, + "grad_norm": 3.765625, + "learning_rate": 1.8131342418092202e-05, + "loss": 0.954, + "step": 1290 + }, + { + "epoch": 0.1884672538146497, + "grad_norm": 4.65625, + "learning_rate": 1.8116845462452886e-05, + "loss": 0.9436, + "step": 1300 + }, + { + "epoch": 0.18991700192091623, + "grad_norm": 3.953125, + "learning_rate": 1.810234850681357e-05, + "loss": 0.9598, + "step": 1310 + }, + { + "epoch": 0.19136675002718279, + "grad_norm": 4.375, + "learning_rate": 1.8087851551174255e-05, + "loss": 0.9772, + "step": 1320 + }, + { + "epoch": 0.1928164981334493, + "grad_norm": 4.25, + "learning_rate": 1.807335459553494e-05, + "loss": 0.9666, + "step": 1330 + }, + { + "epoch": 0.19426624623971586, + "grad_norm": 4.4375, + "learning_rate": 1.8058857639895624e-05, + "loss": 0.9915, + "step": 1340 + }, + { + "epoch": 0.19571599434598239, + "grad_norm": 3.65625, + "learning_rate": 1.804436068425631e-05, + "loss": 0.9198, + "step": 1350 + }, + { + "epoch": 0.1971657424522489, + "grad_norm": 4.28125, + "learning_rate": 1.8029863728616993e-05, + "loss": 0.9415, + "step": 1360 + }, + { + "epoch": 0.19861549055851546, + "grad_norm": 4.5, + "learning_rate": 1.8015366772977677e-05, + "loss": 0.8826, + "step": 1370 + }, + { + "epoch": 0.20006523866478199, + "grad_norm": 4.375, + "learning_rate": 1.8000869817338362e-05, + "loss": 0.9681, + "step": 1380 + }, + { + "epoch": 0.20151498677104854, + "grad_norm": 3.890625, + "learning_rate": 1.7986372861699043e-05, + "loss": 0.9526, + "step": 1390 + }, + { + "epoch": 0.20296473487731506, + "grad_norm": 3.625, + "learning_rate": 1.7971875906059727e-05, + "loss": 0.9296, + "step": 1400 + }, + { + "epoch": 0.2044144829835816, + "grad_norm": 3.5, + "learning_rate": 1.795737895042041e-05, + "loss": 0.9136, + "step": 1410 + }, + { + "epoch": 0.20586423108984814, + "grad_norm": 3.578125, + "learning_rate": 1.79428819947811e-05, + "loss": 0.9171, + "step": 1420 + }, + { + "epoch": 0.20731397919611466, + "grad_norm": 19.75, + "learning_rate": 1.7928385039141784e-05, + "loss": 0.8676, + "step": 1430 + }, + { + "epoch": 0.2087637273023812, + "grad_norm": 3.34375, + "learning_rate": 1.7913888083502465e-05, + "loss": 0.8419, + "step": 1440 + }, + { + "epoch": 0.21021347540864774, + "grad_norm": 4.96875, + "learning_rate": 1.789939112786315e-05, + "loss": 0.8553, + "step": 1450 + }, + { + "epoch": 0.2116632235149143, + "grad_norm": 3.703125, + "learning_rate": 1.7884894172223834e-05, + "loss": 0.8858, + "step": 1460 + }, + { + "epoch": 0.2131129716211808, + "grad_norm": 9.875, + "learning_rate": 1.7870397216584518e-05, + "loss": 0.8399, + "step": 1470 + }, + { + "epoch": 0.21456271972744737, + "grad_norm": 2.96875, + "learning_rate": 1.7855900260945202e-05, + "loss": 0.8398, + "step": 1480 + }, + { + "epoch": 0.2160124678337139, + "grad_norm": 3.515625, + "learning_rate": 1.7841403305305887e-05, + "loss": 0.8399, + "step": 1490 + }, + { + "epoch": 0.21746221593998044, + "grad_norm": 3.15625, + "learning_rate": 1.782690634966657e-05, + "loss": 0.7819, + "step": 1500 + }, + { + "epoch": 0.21891196404624697, + "grad_norm": 3.40625, + "learning_rate": 1.7812409394027256e-05, + "loss": 0.845, + "step": 1510 + }, + { + "epoch": 0.2203617121525135, + "grad_norm": 3.015625, + "learning_rate": 1.779791243838794e-05, + "loss": 0.8055, + "step": 1520 + }, + { + "epoch": 0.22181146025878004, + "grad_norm": 3.140625, + "learning_rate": 1.7783415482748625e-05, + "loss": 0.8237, + "step": 1530 + }, + { + "epoch": 0.22326120836504657, + "grad_norm": 3.15625, + "learning_rate": 1.776891852710931e-05, + "loss": 0.8217, + "step": 1540 + }, + { + "epoch": 0.22471095647131312, + "grad_norm": 2.859375, + "learning_rate": 1.7754421571469993e-05, + "loss": 0.815, + "step": 1550 + }, + { + "epoch": 0.22616070457757964, + "grad_norm": 2.953125, + "learning_rate": 1.7739924615830678e-05, + "loss": 0.8116, + "step": 1560 + }, + { + "epoch": 0.2276104526838462, + "grad_norm": 2.859375, + "learning_rate": 1.7725427660191362e-05, + "loss": 0.8284, + "step": 1570 + }, + { + "epoch": 0.22906020079011272, + "grad_norm": 3.65625, + "learning_rate": 1.7710930704552047e-05, + "loss": 0.8, + "step": 1580 + }, + { + "epoch": 0.23050994889637924, + "grad_norm": 2.875, + "learning_rate": 1.7696433748912728e-05, + "loss": 0.7552, + "step": 1590 + }, + { + "epoch": 0.2319596970026458, + "grad_norm": 2.875, + "learning_rate": 1.7681936793273412e-05, + "loss": 0.7149, + "step": 1600 + }, + { + "epoch": 0.23340944510891232, + "grad_norm": 3.171875, + "learning_rate": 1.7667439837634097e-05, + "loss": 0.7254, + "step": 1610 + }, + { + "epoch": 0.23485919321517887, + "grad_norm": 3.25, + "learning_rate": 1.7652942881994784e-05, + "loss": 0.7839, + "step": 1620 + }, + { + "epoch": 0.2363089413214454, + "grad_norm": 3.265625, + "learning_rate": 1.763844592635547e-05, + "loss": 0.7774, + "step": 1630 + }, + { + "epoch": 0.23775868942771194, + "grad_norm": 3.1875, + "learning_rate": 1.7623948970716153e-05, + "loss": 0.771, + "step": 1640 + }, + { + "epoch": 0.23920843753397847, + "grad_norm": 2.859375, + "learning_rate": 1.7609452015076834e-05, + "loss": 0.748, + "step": 1650 + }, + { + "epoch": 0.24065818564024502, + "grad_norm": 3.09375, + "learning_rate": 1.759495505943752e-05, + "loss": 0.7521, + "step": 1660 + }, + { + "epoch": 0.24210793374651154, + "grad_norm": 3.140625, + "learning_rate": 1.7580458103798203e-05, + "loss": 0.7102, + "step": 1670 + }, + { + "epoch": 0.24355768185277807, + "grad_norm": 2.984375, + "learning_rate": 1.7565961148158887e-05, + "loss": 0.7081, + "step": 1680 + }, + { + "epoch": 0.24500742995904462, + "grad_norm": 3.28125, + "learning_rate": 1.7551464192519572e-05, + "loss": 0.7104, + "step": 1690 + }, + { + "epoch": 0.24645717806531114, + "grad_norm": 3.28125, + "learning_rate": 1.7536967236880256e-05, + "loss": 0.6859, + "step": 1700 + }, + { + "epoch": 0.2479069261715777, + "grad_norm": 3.046875, + "learning_rate": 1.752247028124094e-05, + "loss": 0.6955, + "step": 1710 + }, + { + "epoch": 0.24935667427784422, + "grad_norm": 3.546875, + "learning_rate": 1.7507973325601625e-05, + "loss": 0.7151, + "step": 1720 + }, + { + "epoch": 0.25080642238411077, + "grad_norm": 2.859375, + "learning_rate": 1.749347636996231e-05, + "loss": 0.74, + "step": 1730 + }, + { + "epoch": 0.2522561704903773, + "grad_norm": 2.546875, + "learning_rate": 1.7478979414322994e-05, + "loss": 0.6606, + "step": 1740 + }, + { + "epoch": 0.2537059185966438, + "grad_norm": 2.765625, + "learning_rate": 1.746448245868368e-05, + "loss": 0.6891, + "step": 1750 + }, + { + "epoch": 0.25515566670291034, + "grad_norm": 2.953125, + "learning_rate": 1.7449985503044363e-05, + "loss": 0.7079, + "step": 1760 + }, + { + "epoch": 0.2566054148091769, + "grad_norm": 3.03125, + "learning_rate": 1.7435488547405047e-05, + "loss": 0.7503, + "step": 1770 + }, + { + "epoch": 0.25805516291544345, + "grad_norm": 3.265625, + "learning_rate": 1.742099159176573e-05, + "loss": 0.6945, + "step": 1780 + }, + { + "epoch": 0.25950491102170997, + "grad_norm": 2.84375, + "learning_rate": 1.7406494636126413e-05, + "loss": 0.6704, + "step": 1790 + }, + { + "epoch": 0.2609546591279765, + "grad_norm": 3.0, + "learning_rate": 1.7391997680487097e-05, + "loss": 0.6345, + "step": 1800 + }, + { + "epoch": 0.2624044072342431, + "grad_norm": 3.0, + "learning_rate": 1.7377500724847785e-05, + "loss": 0.6334, + "step": 1810 + }, + { + "epoch": 0.2638541553405096, + "grad_norm": 2.96875, + "learning_rate": 1.736300376920847e-05, + "loss": 0.6747, + "step": 1820 + }, + { + "epoch": 0.2653039034467761, + "grad_norm": 3.015625, + "learning_rate": 1.7348506813569154e-05, + "loss": 0.6682, + "step": 1830 + }, + { + "epoch": 0.26675365155304265, + "grad_norm": 2.96875, + "learning_rate": 1.7334009857929838e-05, + "loss": 0.6314, + "step": 1840 + }, + { + "epoch": 0.26820339965930917, + "grad_norm": 3.1875, + "learning_rate": 1.731951290229052e-05, + "loss": 0.6588, + "step": 1850 + }, + { + "epoch": 0.26965314776557575, + "grad_norm": 2.59375, + "learning_rate": 1.7305015946651204e-05, + "loss": 0.663, + "step": 1860 + }, + { + "epoch": 0.2711028958718423, + "grad_norm": 2.921875, + "learning_rate": 1.7290518991011888e-05, + "loss": 0.6683, + "step": 1870 + }, + { + "epoch": 0.2725526439781088, + "grad_norm": 2.875, + "learning_rate": 1.7276022035372572e-05, + "loss": 0.6402, + "step": 1880 + }, + { + "epoch": 0.2740023920843753, + "grad_norm": 2.75, + "learning_rate": 1.7261525079733257e-05, + "loss": 0.6216, + "step": 1890 + }, + { + "epoch": 0.2754521401906419, + "grad_norm": 2.9375, + "learning_rate": 1.724702812409394e-05, + "loss": 0.6139, + "step": 1900 + }, + { + "epoch": 0.2769018882969084, + "grad_norm": 3.328125, + "learning_rate": 1.7232531168454626e-05, + "loss": 0.6549, + "step": 1910 + }, + { + "epoch": 0.27835163640317495, + "grad_norm": 2.96875, + "learning_rate": 1.721803421281531e-05, + "loss": 0.5732, + "step": 1920 + }, + { + "epoch": 0.2798013845094415, + "grad_norm": 2.640625, + "learning_rate": 1.7203537257175994e-05, + "loss": 0.6259, + "step": 1930 + }, + { + "epoch": 0.281251132615708, + "grad_norm": 2.734375, + "learning_rate": 1.718904030153668e-05, + "loss": 0.63, + "step": 1940 + }, + { + "epoch": 0.2827008807219746, + "grad_norm": 2.78125, + "learning_rate": 1.7174543345897363e-05, + "loss": 0.6002, + "step": 1950 + }, + { + "epoch": 0.2841506288282411, + "grad_norm": 2.96875, + "learning_rate": 1.7160046390258048e-05, + "loss": 0.6192, + "step": 1960 + }, + { + "epoch": 0.2856003769345076, + "grad_norm": 3.125, + "learning_rate": 1.7145549434618732e-05, + "loss": 0.5734, + "step": 1970 + }, + { + "epoch": 0.28705012504077415, + "grad_norm": 2.953125, + "learning_rate": 1.7131052478979417e-05, + "loss": 0.5924, + "step": 1980 + }, + { + "epoch": 0.2884998731470407, + "grad_norm": 2.859375, + "learning_rate": 1.71165555233401e-05, + "loss": 0.5542, + "step": 1990 + }, + { + "epoch": 0.28994962125330725, + "grad_norm": 2.703125, + "learning_rate": 1.7102058567700782e-05, + "loss": 0.5508, + "step": 2000 + }, + { + "epoch": 0.2913993693595738, + "grad_norm": 2.5, + "learning_rate": 1.708756161206147e-05, + "loss": 0.5542, + "step": 2010 + }, + { + "epoch": 0.2928491174658403, + "grad_norm": 2.90625, + "learning_rate": 1.7073064656422154e-05, + "loss": 0.564, + "step": 2020 + }, + { + "epoch": 0.2942988655721068, + "grad_norm": 2.84375, + "learning_rate": 1.705856770078284e-05, + "loss": 0.5562, + "step": 2030 + }, + { + "epoch": 0.2957486136783734, + "grad_norm": 2.953125, + "learning_rate": 1.7044070745143523e-05, + "loss": 0.608, + "step": 2040 + }, + { + "epoch": 0.29719836178463993, + "grad_norm": 2.421875, + "learning_rate": 1.7029573789504204e-05, + "loss": 0.559, + "step": 2050 + }, + { + "epoch": 0.29864810989090645, + "grad_norm": 2.828125, + "learning_rate": 1.701507683386489e-05, + "loss": 0.594, + "step": 2060 + }, + { + "epoch": 0.300097857997173, + "grad_norm": 2.703125, + "learning_rate": 1.7000579878225573e-05, + "loss": 0.5499, + "step": 2070 + }, + { + "epoch": 0.3015476061034395, + "grad_norm": 2.671875, + "learning_rate": 1.6986082922586257e-05, + "loss": 0.5656, + "step": 2080 + }, + { + "epoch": 0.3029973542097061, + "grad_norm": 2.625, + "learning_rate": 1.6971585966946942e-05, + "loss": 0.552, + "step": 2090 + }, + { + "epoch": 0.3044471023159726, + "grad_norm": 3.046875, + "learning_rate": 1.6957089011307626e-05, + "loss": 0.606, + "step": 2100 + }, + { + "epoch": 0.30589685042223913, + "grad_norm": 2.78125, + "learning_rate": 1.694259205566831e-05, + "loss": 0.5693, + "step": 2110 + }, + { + "epoch": 0.30734659852850565, + "grad_norm": 3.1875, + "learning_rate": 1.6928095100028995e-05, + "loss": 0.5844, + "step": 2120 + }, + { + "epoch": 0.30879634663477223, + "grad_norm": 2.71875, + "learning_rate": 1.691359814438968e-05, + "loss": 0.5218, + "step": 2130 + }, + { + "epoch": 0.31024609474103876, + "grad_norm": 2.765625, + "learning_rate": 1.6899101188750364e-05, + "loss": 0.5425, + "step": 2140 + }, + { + "epoch": 0.3116958428473053, + "grad_norm": 2.796875, + "learning_rate": 1.6884604233111048e-05, + "loss": 0.5616, + "step": 2150 + }, + { + "epoch": 0.3131455909535718, + "grad_norm": 2.828125, + "learning_rate": 1.6870107277471733e-05, + "loss": 0.557, + "step": 2160 + }, + { + "epoch": 0.31459533905983833, + "grad_norm": 2.71875, + "learning_rate": 1.6855610321832417e-05, + "loss": 0.5053, + "step": 2170 + }, + { + "epoch": 0.3160450871661049, + "grad_norm": 3.078125, + "learning_rate": 1.68411133661931e-05, + "loss": 0.5129, + "step": 2180 + }, + { + "epoch": 0.31749483527237143, + "grad_norm": 2.546875, + "learning_rate": 1.6826616410553786e-05, + "loss": 0.5499, + "step": 2190 + }, + { + "epoch": 0.31894458337863796, + "grad_norm": 2.4375, + "learning_rate": 1.6812119454914467e-05, + "loss": 0.5033, + "step": 2200 + }, + { + "epoch": 0.3203943314849045, + "grad_norm": 2.84375, + "learning_rate": 1.6797622499275155e-05, + "loss": 0.5304, + "step": 2210 + }, + { + "epoch": 0.32184407959117106, + "grad_norm": 2.921875, + "learning_rate": 1.678312554363584e-05, + "loss": 0.5024, + "step": 2220 + }, + { + "epoch": 0.3232938276974376, + "grad_norm": 2.5625, + "learning_rate": 1.6768628587996524e-05, + "loss": 0.497, + "step": 2230 + }, + { + "epoch": 0.3247435758037041, + "grad_norm": 2.640625, + "learning_rate": 1.6754131632357208e-05, + "loss": 0.5155, + "step": 2240 + }, + { + "epoch": 0.32619332390997063, + "grad_norm": 2.671875, + "learning_rate": 1.6739634676717892e-05, + "loss": 0.5126, + "step": 2250 + }, + { + "epoch": 0.32764307201623716, + "grad_norm": 2.78125, + "learning_rate": 1.6725137721078573e-05, + "loss": 0.5221, + "step": 2260 + }, + { + "epoch": 0.32909282012250374, + "grad_norm": 2.765625, + "learning_rate": 1.6710640765439258e-05, + "loss": 0.5153, + "step": 2270 + }, + { + "epoch": 0.33054256822877026, + "grad_norm": 2.9375, + "learning_rate": 1.6696143809799942e-05, + "loss": 0.5342, + "step": 2280 + }, + { + "epoch": 0.3319923163350368, + "grad_norm": 2.625, + "learning_rate": 1.6681646854160627e-05, + "loss": 0.4855, + "step": 2290 + }, + { + "epoch": 0.3334420644413033, + "grad_norm": 2.78125, + "learning_rate": 1.666714989852131e-05, + "loss": 0.4692, + "step": 2300 + }, + { + "epoch": 0.33489181254756983, + "grad_norm": 2.515625, + "learning_rate": 1.6652652942881996e-05, + "loss": 0.4713, + "step": 2310 + }, + { + "epoch": 0.3363415606538364, + "grad_norm": 2.640625, + "learning_rate": 1.663815598724268e-05, + "loss": 0.4787, + "step": 2320 + }, + { + "epoch": 0.33779130876010294, + "grad_norm": 2.734375, + "learning_rate": 1.6623659031603364e-05, + "loss": 0.4909, + "step": 2330 + }, + { + "epoch": 0.33924105686636946, + "grad_norm": 2.734375, + "learning_rate": 1.660916207596405e-05, + "loss": 0.4572, + "step": 2340 + }, + { + "epoch": 0.340690804972636, + "grad_norm": 2.359375, + "learning_rate": 1.6594665120324733e-05, + "loss": 0.4744, + "step": 2350 + }, + { + "epoch": 0.34214055307890257, + "grad_norm": 2.734375, + "learning_rate": 1.6580168164685418e-05, + "loss": 0.4718, + "step": 2360 + }, + { + "epoch": 0.3435903011851691, + "grad_norm": 2.734375, + "learning_rate": 1.6565671209046102e-05, + "loss": 0.4431, + "step": 2370 + }, + { + "epoch": 0.3450400492914356, + "grad_norm": 2.796875, + "learning_rate": 1.6551174253406786e-05, + "loss": 0.4819, + "step": 2380 + }, + { + "epoch": 0.34648979739770214, + "grad_norm": 3.09375, + "learning_rate": 1.653667729776747e-05, + "loss": 0.4923, + "step": 2390 + }, + { + "epoch": 0.34793954550396866, + "grad_norm": 2.953125, + "learning_rate": 1.6522180342128152e-05, + "loss": 0.4772, + "step": 2400 + }, + { + "epoch": 0.34938929361023524, + "grad_norm": 2.515625, + "learning_rate": 1.650768338648884e-05, + "loss": 0.4737, + "step": 2410 + }, + { + "epoch": 0.35083904171650176, + "grad_norm": 3.3125, + "learning_rate": 1.6493186430849524e-05, + "loss": 0.4902, + "step": 2420 + }, + { + "epoch": 0.3522887898227683, + "grad_norm": 3.515625, + "learning_rate": 1.647868947521021e-05, + "loss": 0.4489, + "step": 2430 + }, + { + "epoch": 0.3537385379290348, + "grad_norm": 2.484375, + "learning_rate": 1.6464192519570893e-05, + "loss": 0.4554, + "step": 2440 + }, + { + "epoch": 0.3551882860353014, + "grad_norm": 2.6875, + "learning_rate": 1.6449695563931577e-05, + "loss": 0.4957, + "step": 2450 + }, + { + "epoch": 0.3566380341415679, + "grad_norm": 2.765625, + "learning_rate": 1.643519860829226e-05, + "loss": 0.439, + "step": 2460 + }, + { + "epoch": 0.35808778224783444, + "grad_norm": 2.703125, + "learning_rate": 1.6420701652652943e-05, + "loss": 0.4508, + "step": 2470 + }, + { + "epoch": 0.35953753035410096, + "grad_norm": 2.625, + "learning_rate": 1.6406204697013627e-05, + "loss": 0.4258, + "step": 2480 + }, + { + "epoch": 0.3609872784603675, + "grad_norm": 2.71875, + "learning_rate": 1.639170774137431e-05, + "loss": 0.4699, + "step": 2490 + }, + { + "epoch": 0.36243702656663407, + "grad_norm": 2.578125, + "learning_rate": 1.6377210785734996e-05, + "loss": 0.432, + "step": 2500 + }, + { + "epoch": 0.3638867746729006, + "grad_norm": 2.3125, + "learning_rate": 1.6362713830095684e-05, + "loss": 0.4304, + "step": 2510 + }, + { + "epoch": 0.3653365227791671, + "grad_norm": 2.78125, + "learning_rate": 1.6348216874456365e-05, + "loss": 0.4805, + "step": 2520 + }, + { + "epoch": 0.36678627088543364, + "grad_norm": 2.453125, + "learning_rate": 1.633371991881705e-05, + "loss": 0.4604, + "step": 2530 + }, + { + "epoch": 0.36823601899170016, + "grad_norm": 2.265625, + "learning_rate": 1.6319222963177734e-05, + "loss": 0.4803, + "step": 2540 + }, + { + "epoch": 0.36968576709796674, + "grad_norm": 2.65625, + "learning_rate": 1.6304726007538418e-05, + "loss": 0.4259, + "step": 2550 + }, + { + "epoch": 0.37113551520423327, + "grad_norm": 2.890625, + "learning_rate": 1.6290229051899103e-05, + "loss": 0.4283, + "step": 2560 + }, + { + "epoch": 0.3725852633104998, + "grad_norm": 2.65625, + "learning_rate": 1.6275732096259787e-05, + "loss": 0.4135, + "step": 2570 + }, + { + "epoch": 0.3740350114167663, + "grad_norm": 2.625, + "learning_rate": 1.626123514062047e-05, + "loss": 0.4054, + "step": 2580 + }, + { + "epoch": 0.3754847595230329, + "grad_norm": 2.890625, + "learning_rate": 1.6246738184981156e-05, + "loss": 0.3739, + "step": 2590 + }, + { + "epoch": 0.3769345076292994, + "grad_norm": 2.453125, + "learning_rate": 1.623224122934184e-05, + "loss": 0.417, + "step": 2600 + }, + { + "epoch": 0.37838425573556594, + "grad_norm": 2.671875, + "learning_rate": 1.6217744273702525e-05, + "loss": 0.3945, + "step": 2610 + }, + { + "epoch": 0.37983400384183247, + "grad_norm": 2.515625, + "learning_rate": 1.620324731806321e-05, + "loss": 0.4265, + "step": 2620 + }, + { + "epoch": 0.381283751948099, + "grad_norm": 2.34375, + "learning_rate": 1.6188750362423893e-05, + "loss": 0.4437, + "step": 2630 + }, + { + "epoch": 0.38273350005436557, + "grad_norm": 2.8125, + "learning_rate": 1.6174253406784578e-05, + "loss": 0.4536, + "step": 2640 + }, + { + "epoch": 0.3841832481606321, + "grad_norm": 2.96875, + "learning_rate": 1.6159756451145262e-05, + "loss": 0.4168, + "step": 2650 + }, + { + "epoch": 0.3856329962668986, + "grad_norm": 2.578125, + "learning_rate": 1.6145259495505943e-05, + "loss": 0.4287, + "step": 2660 + }, + { + "epoch": 0.38708274437316514, + "grad_norm": 2.9375, + "learning_rate": 1.6130762539866628e-05, + "loss": 0.433, + "step": 2670 + }, + { + "epoch": 0.3885324924794317, + "grad_norm": 2.703125, + "learning_rate": 1.6116265584227312e-05, + "loss": 0.4092, + "step": 2680 + }, + { + "epoch": 0.38998224058569825, + "grad_norm": 2.390625, + "learning_rate": 1.6101768628587997e-05, + "loss": 0.3932, + "step": 2690 + }, + { + "epoch": 0.39143198869196477, + "grad_norm": 2.703125, + "learning_rate": 1.608727167294868e-05, + "loss": 0.4219, + "step": 2700 + }, + { + "epoch": 0.3928817367982313, + "grad_norm": 2.828125, + "learning_rate": 1.607277471730937e-05, + "loss": 0.4117, + "step": 2710 + }, + { + "epoch": 0.3943314849044978, + "grad_norm": 2.78125, + "learning_rate": 1.605827776167005e-05, + "loss": 0.3557, + "step": 2720 + }, + { + "epoch": 0.3957812330107644, + "grad_norm": 2.375, + "learning_rate": 1.6043780806030734e-05, + "loss": 0.3723, + "step": 2730 + }, + { + "epoch": 0.3972309811170309, + "grad_norm": 3.09375, + "learning_rate": 1.602928385039142e-05, + "loss": 0.3992, + "step": 2740 + }, + { + "epoch": 0.39868072922329745, + "grad_norm": 2.84375, + "learning_rate": 1.6014786894752103e-05, + "loss": 0.4101, + "step": 2750 + }, + { + "epoch": 0.40013047732956397, + "grad_norm": 2.40625, + "learning_rate": 1.6000289939112787e-05, + "loss": 0.4028, + "step": 2760 + }, + { + "epoch": 0.40158022543583055, + "grad_norm": 2.75, + "learning_rate": 1.5985792983473472e-05, + "loss": 0.4268, + "step": 2770 + }, + { + "epoch": 0.4030299735420971, + "grad_norm": 2.34375, + "learning_rate": 1.5971296027834156e-05, + "loss": 0.382, + "step": 2780 + }, + { + "epoch": 0.4044797216483636, + "grad_norm": 2.640625, + "learning_rate": 1.595679907219484e-05, + "loss": 0.4143, + "step": 2790 + }, + { + "epoch": 0.4059294697546301, + "grad_norm": 2.375, + "learning_rate": 1.5942302116555525e-05, + "loss": 0.3776, + "step": 2800 + }, + { + "epoch": 0.40737921786089665, + "grad_norm": 2.609375, + "learning_rate": 1.592780516091621e-05, + "loss": 0.3881, + "step": 2810 + }, + { + "epoch": 0.4088289659671632, + "grad_norm": 2.453125, + "learning_rate": 1.5913308205276894e-05, + "loss": 0.3945, + "step": 2820 + }, + { + "epoch": 0.41027871407342975, + "grad_norm": 2.21875, + "learning_rate": 1.589881124963758e-05, + "loss": 0.3863, + "step": 2830 + }, + { + "epoch": 0.4117284621796963, + "grad_norm": 2.40625, + "learning_rate": 1.5884314293998263e-05, + "loss": 0.3503, + "step": 2840 + }, + { + "epoch": 0.4131782102859628, + "grad_norm": 2.703125, + "learning_rate": 1.5869817338358947e-05, + "loss": 0.3564, + "step": 2850 + }, + { + "epoch": 0.4146279583922293, + "grad_norm": 2.484375, + "learning_rate": 1.585532038271963e-05, + "loss": 0.3654, + "step": 2860 + }, + { + "epoch": 0.4160777064984959, + "grad_norm": 2.8125, + "learning_rate": 1.5840823427080313e-05, + "loss": 0.41, + "step": 2870 + }, + { + "epoch": 0.4175274546047624, + "grad_norm": 2.65625, + "learning_rate": 1.5826326471440997e-05, + "loss": 0.3976, + "step": 2880 + }, + { + "epoch": 0.41897720271102895, + "grad_norm": 2.4375, + "learning_rate": 1.581182951580168e-05, + "loss": 0.3643, + "step": 2890 + }, + { + "epoch": 0.4204269508172955, + "grad_norm": 2.40625, + "learning_rate": 1.5797332560162366e-05, + "loss": 0.3715, + "step": 2900 + }, + { + "epoch": 0.42187669892356205, + "grad_norm": 2.59375, + "learning_rate": 1.5782835604523054e-05, + "loss": 0.3685, + "step": 2910 + }, + { + "epoch": 0.4233264470298286, + "grad_norm": 2.3125, + "learning_rate": 1.5768338648883735e-05, + "loss": 0.3913, + "step": 2920 + }, + { + "epoch": 0.4247761951360951, + "grad_norm": 2.484375, + "learning_rate": 1.575384169324442e-05, + "loss": 0.3755, + "step": 2930 + }, + { + "epoch": 0.4262259432423616, + "grad_norm": 2.21875, + "learning_rate": 1.5739344737605104e-05, + "loss": 0.3715, + "step": 2940 + }, + { + "epoch": 0.42767569134862815, + "grad_norm": 2.15625, + "learning_rate": 1.5724847781965788e-05, + "loss": 0.366, + "step": 2950 + }, + { + "epoch": 0.42912543945489473, + "grad_norm": 2.90625, + "learning_rate": 1.5710350826326472e-05, + "loss": 0.3655, + "step": 2960 + }, + { + "epoch": 0.43057518756116125, + "grad_norm": 2.59375, + "learning_rate": 1.5695853870687157e-05, + "loss": 0.3745, + "step": 2970 + }, + { + "epoch": 0.4320249356674278, + "grad_norm": 2.71875, + "learning_rate": 1.568135691504784e-05, + "loss": 0.3592, + "step": 2980 + }, + { + "epoch": 0.4334746837736943, + "grad_norm": 2.328125, + "learning_rate": 1.5666859959408526e-05, + "loss": 0.3933, + "step": 2990 + }, + { + "epoch": 0.4349244318799609, + "grad_norm": 2.46875, + "learning_rate": 1.565236300376921e-05, + "loss": 0.3456, + "step": 3000 + }, + { + "epoch": 0.4363741799862274, + "grad_norm": 2.53125, + "learning_rate": 1.5637866048129895e-05, + "loss": 0.343, + "step": 3010 + }, + { + "epoch": 0.43782392809249393, + "grad_norm": 2.296875, + "learning_rate": 1.562336909249058e-05, + "loss": 0.3819, + "step": 3020 + }, + { + "epoch": 0.43927367619876045, + "grad_norm": 2.375, + "learning_rate": 1.5608872136851263e-05, + "loss": 0.3599, + "step": 3030 + }, + { + "epoch": 0.440723424305027, + "grad_norm": 2.078125, + "learning_rate": 1.5594375181211948e-05, + "loss": 0.3384, + "step": 3040 + }, + { + "epoch": 0.44217317241129356, + "grad_norm": 2.375, + "learning_rate": 1.5579878225572632e-05, + "loss": 0.3701, + "step": 3050 + }, + { + "epoch": 0.4436229205175601, + "grad_norm": 2.828125, + "learning_rate": 1.5565381269933317e-05, + "loss": 0.3403, + "step": 3060 + }, + { + "epoch": 0.4450726686238266, + "grad_norm": 2.328125, + "learning_rate": 1.5550884314293998e-05, + "loss": 0.3721, + "step": 3070 + }, + { + "epoch": 0.44652241673009313, + "grad_norm": 2.640625, + "learning_rate": 1.5536387358654682e-05, + "loss": 0.3634, + "step": 3080 + }, + { + "epoch": 0.4479721648363597, + "grad_norm": 2.59375, + "learning_rate": 1.5521890403015366e-05, + "loss": 0.3548, + "step": 3090 + }, + { + "epoch": 0.44942191294262623, + "grad_norm": 2.328125, + "learning_rate": 1.550739344737605e-05, + "loss": 0.3467, + "step": 3100 + }, + { + "epoch": 0.45087166104889276, + "grad_norm": 2.484375, + "learning_rate": 1.549289649173674e-05, + "loss": 0.3681, + "step": 3110 + }, + { + "epoch": 0.4523214091551593, + "grad_norm": 2.203125, + "learning_rate": 1.5478399536097423e-05, + "loss": 0.3505, + "step": 3120 + }, + { + "epoch": 0.4537711572614258, + "grad_norm": 2.734375, + "learning_rate": 1.5463902580458104e-05, + "loss": 0.3434, + "step": 3130 + }, + { + "epoch": 0.4552209053676924, + "grad_norm": 2.1875, + "learning_rate": 1.544940562481879e-05, + "loss": 0.3699, + "step": 3140 + }, + { + "epoch": 0.4566706534739589, + "grad_norm": 2.46875, + "learning_rate": 1.5434908669179473e-05, + "loss": 0.3535, + "step": 3150 + }, + { + "epoch": 0.45812040158022543, + "grad_norm": 2.375, + "learning_rate": 1.5420411713540157e-05, + "loss": 0.347, + "step": 3160 + }, + { + "epoch": 0.45957014968649196, + "grad_norm": 2.4375, + "learning_rate": 1.5405914757900842e-05, + "loss": 0.3246, + "step": 3170 + }, + { + "epoch": 0.4610198977927585, + "grad_norm": 2.359375, + "learning_rate": 1.5391417802261526e-05, + "loss": 0.3855, + "step": 3180 + }, + { + "epoch": 0.46246964589902506, + "grad_norm": 2.703125, + "learning_rate": 1.537692084662221e-05, + "loss": 0.337, + "step": 3190 + }, + { + "epoch": 0.4639193940052916, + "grad_norm": 2.09375, + "learning_rate": 1.5362423890982895e-05, + "loss": 0.3662, + "step": 3200 + }, + { + "epoch": 0.4653691421115581, + "grad_norm": 2.421875, + "learning_rate": 1.534792693534358e-05, + "loss": 0.3239, + "step": 3210 + }, + { + "epoch": 0.46681889021782463, + "grad_norm": 2.328125, + "learning_rate": 1.5333429979704264e-05, + "loss": 0.3215, + "step": 3220 + }, + { + "epoch": 0.4682686383240912, + "grad_norm": 2.484375, + "learning_rate": 1.5318933024064948e-05, + "loss": 0.3532, + "step": 3230 + }, + { + "epoch": 0.46971838643035774, + "grad_norm": 2.1875, + "learning_rate": 1.5304436068425633e-05, + "loss": 0.3381, + "step": 3240 + }, + { + "epoch": 0.47116813453662426, + "grad_norm": 2.265625, + "learning_rate": 1.5289939112786317e-05, + "loss": 0.3519, + "step": 3250 + }, + { + "epoch": 0.4726178826428908, + "grad_norm": 2.234375, + "learning_rate": 1.5275442157147e-05, + "loss": 0.3229, + "step": 3260 + }, + { + "epoch": 0.4740676307491573, + "grad_norm": 2.203125, + "learning_rate": 1.5260945201507683e-05, + "loss": 0.3448, + "step": 3270 + }, + { + "epoch": 0.4755173788554239, + "grad_norm": 2.25, + "learning_rate": 1.5246448245868369e-05, + "loss": 0.3678, + "step": 3280 + }, + { + "epoch": 0.4769671269616904, + "grad_norm": 2.40625, + "learning_rate": 1.5231951290229053e-05, + "loss": 0.3244, + "step": 3290 + }, + { + "epoch": 0.47841687506795694, + "grad_norm": 2.546875, + "learning_rate": 1.5217454334589736e-05, + "loss": 0.3434, + "step": 3300 + }, + { + "epoch": 0.47986662317422346, + "grad_norm": 1.984375, + "learning_rate": 1.5202957378950422e-05, + "loss": 0.2983, + "step": 3310 + }, + { + "epoch": 0.48131637128049004, + "grad_norm": 2.71875, + "learning_rate": 1.5188460423311106e-05, + "loss": 0.3329, + "step": 3320 + }, + { + "epoch": 0.48276611938675656, + "grad_norm": 3.0625, + "learning_rate": 1.517396346767179e-05, + "loss": 0.3534, + "step": 3330 + }, + { + "epoch": 0.4842158674930231, + "grad_norm": 1.7890625, + "learning_rate": 1.5159466512032475e-05, + "loss": 0.3777, + "step": 3340 + }, + { + "epoch": 0.4856656155992896, + "grad_norm": 2.4375, + "learning_rate": 1.5144969556393158e-05, + "loss": 0.3539, + "step": 3350 + }, + { + "epoch": 0.48711536370555614, + "grad_norm": 2.46875, + "learning_rate": 1.5130472600753842e-05, + "loss": 0.2904, + "step": 3360 + }, + { + "epoch": 0.4885651118118227, + "grad_norm": 2.71875, + "learning_rate": 1.5115975645114527e-05, + "loss": 0.3442, + "step": 3370 + }, + { + "epoch": 0.49001485991808924, + "grad_norm": 2.25, + "learning_rate": 1.5101478689475211e-05, + "loss": 0.295, + "step": 3380 + }, + { + "epoch": 0.49146460802435576, + "grad_norm": 2.0625, + "learning_rate": 1.5086981733835896e-05, + "loss": 0.3025, + "step": 3390 + }, + { + "epoch": 0.4929143561306223, + "grad_norm": 2.421875, + "learning_rate": 1.5072484778196578e-05, + "loss": 0.3171, + "step": 3400 + }, + { + "epoch": 0.49436410423688887, + "grad_norm": 2.53125, + "learning_rate": 1.5057987822557264e-05, + "loss": 0.3019, + "step": 3410 + }, + { + "epoch": 0.4958138523431554, + "grad_norm": 1.9140625, + "learning_rate": 1.5043490866917949e-05, + "loss": 0.3502, + "step": 3420 + }, + { + "epoch": 0.4972636004494219, + "grad_norm": 2.109375, + "learning_rate": 1.5028993911278633e-05, + "loss": 0.3451, + "step": 3430 + }, + { + "epoch": 0.49871334855568844, + "grad_norm": 2.421875, + "learning_rate": 1.5014496955639318e-05, + "loss": 0.3151, + "step": 3440 + }, + { + "epoch": 0.500163096661955, + "grad_norm": 2.046875, + "learning_rate": 1.5000000000000002e-05, + "loss": 0.3123, + "step": 3450 + }, + { + "epoch": 0.5016128447682215, + "grad_norm": 2.484375, + "learning_rate": 1.4985503044360685e-05, + "loss": 0.3064, + "step": 3460 + }, + { + "epoch": 0.5030625928744881, + "grad_norm": 2.25, + "learning_rate": 1.497100608872137e-05, + "loss": 0.331, + "step": 3470 + }, + { + "epoch": 0.5045123409807546, + "grad_norm": 2.421875, + "learning_rate": 1.4956509133082054e-05, + "loss": 0.2976, + "step": 3480 + }, + { + "epoch": 0.5059620890870211, + "grad_norm": 2.125, + "learning_rate": 1.4942012177442738e-05, + "loss": 0.3207, + "step": 3490 + }, + { + "epoch": 0.5074118371932876, + "grad_norm": 2.0625, + "learning_rate": 1.4927515221803424e-05, + "loss": 0.2869, + "step": 3500 + }, + { + "epoch": 0.5088615852995542, + "grad_norm": 2.203125, + "learning_rate": 1.4913018266164107e-05, + "loss": 0.3411, + "step": 3510 + }, + { + "epoch": 0.5103113334058207, + "grad_norm": 1.78125, + "learning_rate": 1.4898521310524791e-05, + "loss": 0.264, + "step": 3520 + }, + { + "epoch": 0.5117610815120873, + "grad_norm": 2.453125, + "learning_rate": 1.4884024354885476e-05, + "loss": 0.3192, + "step": 3530 + }, + { + "epoch": 0.5132108296183538, + "grad_norm": 2.109375, + "learning_rate": 1.486952739924616e-05, + "loss": 0.2999, + "step": 3540 + }, + { + "epoch": 0.5146605777246204, + "grad_norm": 2.546875, + "learning_rate": 1.4855030443606845e-05, + "loss": 0.2922, + "step": 3550 + }, + { + "epoch": 0.5161103258308869, + "grad_norm": 2.015625, + "learning_rate": 1.4840533487967527e-05, + "loss": 0.329, + "step": 3560 + }, + { + "epoch": 0.5175600739371534, + "grad_norm": 2.25, + "learning_rate": 1.4826036532328212e-05, + "loss": 0.3185, + "step": 3570 + }, + { + "epoch": 0.5190098220434199, + "grad_norm": 2.296875, + "learning_rate": 1.4811539576688896e-05, + "loss": 0.2912, + "step": 3580 + }, + { + "epoch": 0.5204595701496865, + "grad_norm": 2.03125, + "learning_rate": 1.479704262104958e-05, + "loss": 0.2868, + "step": 3590 + }, + { + "epoch": 0.521909318255953, + "grad_norm": 2.5, + "learning_rate": 1.4782545665410267e-05, + "loss": 0.3219, + "step": 3600 + }, + { + "epoch": 0.5233590663622195, + "grad_norm": 2.171875, + "learning_rate": 1.476804870977095e-05, + "loss": 0.3063, + "step": 3610 + }, + { + "epoch": 0.5248088144684862, + "grad_norm": 1.90625, + "learning_rate": 1.4753551754131634e-05, + "loss": 0.3285, + "step": 3620 + }, + { + "epoch": 0.5262585625747527, + "grad_norm": 2.265625, + "learning_rate": 1.4739054798492318e-05, + "loss": 0.3328, + "step": 3630 + }, + { + "epoch": 0.5277083106810192, + "grad_norm": 2.140625, + "learning_rate": 1.4724557842853003e-05, + "loss": 0.3459, + "step": 3640 + }, + { + "epoch": 0.5291580587872857, + "grad_norm": 2.28125, + "learning_rate": 1.4710060887213687e-05, + "loss": 0.2812, + "step": 3650 + }, + { + "epoch": 0.5306078068935522, + "grad_norm": 1.84375, + "learning_rate": 1.469556393157437e-05, + "loss": 0.3076, + "step": 3660 + }, + { + "epoch": 0.5320575549998188, + "grad_norm": 1.984375, + "learning_rate": 1.4681066975935054e-05, + "loss": 0.2806, + "step": 3670 + }, + { + "epoch": 0.5335073031060853, + "grad_norm": 2.453125, + "learning_rate": 1.4666570020295739e-05, + "loss": 0.2855, + "step": 3680 + }, + { + "epoch": 0.5349570512123518, + "grad_norm": 1.9296875, + "learning_rate": 1.4652073064656423e-05, + "loss": 0.2886, + "step": 3690 + }, + { + "epoch": 0.5364067993186183, + "grad_norm": 1.984375, + "learning_rate": 1.4637576109017109e-05, + "loss": 0.2997, + "step": 3700 + }, + { + "epoch": 0.537856547424885, + "grad_norm": 2.1875, + "learning_rate": 1.4623079153377794e-05, + "loss": 0.3326, + "step": 3710 + }, + { + "epoch": 0.5393062955311515, + "grad_norm": 2.4375, + "learning_rate": 1.4608582197738476e-05, + "loss": 0.3038, + "step": 3720 + }, + { + "epoch": 0.540756043637418, + "grad_norm": 2.171875, + "learning_rate": 1.459408524209916e-05, + "loss": 0.277, + "step": 3730 + }, + { + "epoch": 0.5422057917436846, + "grad_norm": 2.015625, + "learning_rate": 1.4579588286459845e-05, + "loss": 0.2777, + "step": 3740 + }, + { + "epoch": 0.5436555398499511, + "grad_norm": 1.9921875, + "learning_rate": 1.456509133082053e-05, + "loss": 0.2858, + "step": 3750 + }, + { + "epoch": 0.5451052879562176, + "grad_norm": 1.9453125, + "learning_rate": 1.4550594375181212e-05, + "loss": 0.3235, + "step": 3760 + }, + { + "epoch": 0.5465550360624841, + "grad_norm": 2.390625, + "learning_rate": 1.4536097419541897e-05, + "loss": 0.3304, + "step": 3770 + }, + { + "epoch": 0.5480047841687506, + "grad_norm": 2.15625, + "learning_rate": 1.4521600463902581e-05, + "loss": 0.2676, + "step": 3780 + }, + { + "epoch": 0.5494545322750172, + "grad_norm": 2.265625, + "learning_rate": 1.4507103508263265e-05, + "loss": 0.2916, + "step": 3790 + }, + { + "epoch": 0.5509042803812838, + "grad_norm": 2.40625, + "learning_rate": 1.4492606552623952e-05, + "loss": 0.315, + "step": 3800 + }, + { + "epoch": 0.5523540284875503, + "grad_norm": 2.390625, + "learning_rate": 1.4478109596984636e-05, + "loss": 0.3041, + "step": 3810 + }, + { + "epoch": 0.5538037765938169, + "grad_norm": 2.03125, + "learning_rate": 1.4463612641345319e-05, + "loss": 0.2836, + "step": 3820 + }, + { + "epoch": 0.5552535247000834, + "grad_norm": 2.359375, + "learning_rate": 1.4449115685706003e-05, + "loss": 0.2971, + "step": 3830 + }, + { + "epoch": 0.5567032728063499, + "grad_norm": 2.265625, + "learning_rate": 1.4434618730066688e-05, + "loss": 0.2856, + "step": 3840 + }, + { + "epoch": 0.5581530209126164, + "grad_norm": 2.5625, + "learning_rate": 1.4420121774427372e-05, + "loss": 0.2874, + "step": 3850 + }, + { + "epoch": 0.559602769018883, + "grad_norm": 2.171875, + "learning_rate": 1.4405624818788055e-05, + "loss": 0.29, + "step": 3860 + }, + { + "epoch": 0.5610525171251495, + "grad_norm": 1.9609375, + "learning_rate": 1.4391127863148739e-05, + "loss": 0.2661, + "step": 3870 + }, + { + "epoch": 0.562502265231416, + "grad_norm": 2.109375, + "learning_rate": 1.4376630907509424e-05, + "loss": 0.3105, + "step": 3880 + }, + { + "epoch": 0.5639520133376825, + "grad_norm": 2.015625, + "learning_rate": 1.4362133951870108e-05, + "loss": 0.3044, + "step": 3890 + }, + { + "epoch": 0.5654017614439492, + "grad_norm": 1.7734375, + "learning_rate": 1.4347636996230794e-05, + "loss": 0.2948, + "step": 3900 + }, + { + "epoch": 0.5668515095502157, + "grad_norm": 2.03125, + "learning_rate": 1.4333140040591478e-05, + "loss": 0.302, + "step": 3910 + }, + { + "epoch": 0.5683012576564822, + "grad_norm": 2.21875, + "learning_rate": 1.4318643084952161e-05, + "loss": 0.2946, + "step": 3920 + }, + { + "epoch": 0.5697510057627487, + "grad_norm": 1.7890625, + "learning_rate": 1.4304146129312846e-05, + "loss": 0.3095, + "step": 3930 + }, + { + "epoch": 0.5712007538690153, + "grad_norm": 2.5625, + "learning_rate": 1.428964917367353e-05, + "loss": 0.3023, + "step": 3940 + }, + { + "epoch": 0.5726505019752818, + "grad_norm": 2.5, + "learning_rate": 1.4275152218034214e-05, + "loss": 0.2455, + "step": 3950 + }, + { + "epoch": 0.5741002500815483, + "grad_norm": 1.9296875, + "learning_rate": 1.4260655262394897e-05, + "loss": 0.2896, + "step": 3960 + }, + { + "epoch": 0.5755499981878148, + "grad_norm": 1.8671875, + "learning_rate": 1.4246158306755582e-05, + "loss": 0.2595, + "step": 3970 + }, + { + "epoch": 0.5769997462940814, + "grad_norm": 1.984375, + "learning_rate": 1.4231661351116266e-05, + "loss": 0.289, + "step": 3980 + }, + { + "epoch": 0.578449494400348, + "grad_norm": 2.109375, + "learning_rate": 1.421716439547695e-05, + "loss": 0.2666, + "step": 3990 + }, + { + "epoch": 0.5798992425066145, + "grad_norm": 2.125, + "learning_rate": 1.4202667439837637e-05, + "loss": 0.3198, + "step": 4000 + }, + { + "epoch": 0.581348990612881, + "grad_norm": 2.28125, + "learning_rate": 1.4188170484198321e-05, + "loss": 0.2666, + "step": 4010 + }, + { + "epoch": 0.5827987387191476, + "grad_norm": 2.09375, + "learning_rate": 1.4173673528559004e-05, + "loss": 0.2831, + "step": 4020 + }, + { + "epoch": 0.5842484868254141, + "grad_norm": 2.0, + "learning_rate": 1.4159176572919688e-05, + "loss": 0.2414, + "step": 4030 + }, + { + "epoch": 0.5856982349316806, + "grad_norm": 1.90625, + "learning_rate": 1.4144679617280372e-05, + "loss": 0.2973, + "step": 4040 + }, + { + "epoch": 0.5871479830379471, + "grad_norm": 2.203125, + "learning_rate": 1.4130182661641057e-05, + "loss": 0.2725, + "step": 4050 + }, + { + "epoch": 0.5885977311442137, + "grad_norm": 1.7265625, + "learning_rate": 1.4115685706001741e-05, + "loss": 0.2716, + "step": 4060 + }, + { + "epoch": 0.5900474792504802, + "grad_norm": 2.671875, + "learning_rate": 1.4101188750362424e-05, + "loss": 0.2977, + "step": 4070 + }, + { + "epoch": 0.5914972273567468, + "grad_norm": 2.171875, + "learning_rate": 1.4086691794723108e-05, + "loss": 0.251, + "step": 4080 + }, + { + "epoch": 0.5929469754630133, + "grad_norm": 2.4375, + "learning_rate": 1.4072194839083793e-05, + "loss": 0.3201, + "step": 4090 + }, + { + "epoch": 0.5943967235692799, + "grad_norm": 1.9609375, + "learning_rate": 1.4057697883444479e-05, + "loss": 0.2786, + "step": 4100 + }, + { + "epoch": 0.5958464716755464, + "grad_norm": 1.921875, + "learning_rate": 1.4043200927805163e-05, + "loss": 0.2939, + "step": 4110 + }, + { + "epoch": 0.5972962197818129, + "grad_norm": 1.8828125, + "learning_rate": 1.4028703972165846e-05, + "loss": 0.2243, + "step": 4120 + }, + { + "epoch": 0.5987459678880794, + "grad_norm": 1.859375, + "learning_rate": 1.401420701652653e-05, + "loss": 0.2625, + "step": 4130 + }, + { + "epoch": 0.600195715994346, + "grad_norm": 1.953125, + "learning_rate": 1.3999710060887215e-05, + "loss": 0.2566, + "step": 4140 + }, + { + "epoch": 0.6016454641006125, + "grad_norm": 2.1875, + "learning_rate": 1.39852131052479e-05, + "loss": 0.297, + "step": 4150 + }, + { + "epoch": 0.603095212206879, + "grad_norm": 1.8984375, + "learning_rate": 1.3970716149608584e-05, + "loss": 0.2866, + "step": 4160 + }, + { + "epoch": 0.6045449603131456, + "grad_norm": 2.109375, + "learning_rate": 1.3956219193969267e-05, + "loss": 0.2851, + "step": 4170 + }, + { + "epoch": 0.6059947084194122, + "grad_norm": 1.7421875, + "learning_rate": 1.3941722238329951e-05, + "loss": 0.2841, + "step": 4180 + }, + { + "epoch": 0.6074444565256787, + "grad_norm": 2.234375, + "learning_rate": 1.3927225282690635e-05, + "loss": 0.2626, + "step": 4190 + }, + { + "epoch": 0.6088942046319452, + "grad_norm": 1.875, + "learning_rate": 1.3912728327051321e-05, + "loss": 0.2509, + "step": 4200 + }, + { + "epoch": 0.6103439527382117, + "grad_norm": 2.171875, + "learning_rate": 1.3898231371412006e-05, + "loss": 0.2815, + "step": 4210 + }, + { + "epoch": 0.6117937008444783, + "grad_norm": 2.09375, + "learning_rate": 1.3883734415772689e-05, + "loss": 0.282, + "step": 4220 + }, + { + "epoch": 0.6132434489507448, + "grad_norm": 1.828125, + "learning_rate": 1.3869237460133373e-05, + "loss": 0.2869, + "step": 4230 + }, + { + "epoch": 0.6146931970570113, + "grad_norm": 1.765625, + "learning_rate": 1.3854740504494057e-05, + "loss": 0.2436, + "step": 4240 + }, + { + "epoch": 0.6161429451632778, + "grad_norm": 2.015625, + "learning_rate": 1.3840243548854742e-05, + "loss": 0.2385, + "step": 4250 + }, + { + "epoch": 0.6175926932695445, + "grad_norm": 2.171875, + "learning_rate": 1.3825746593215426e-05, + "loss": 0.2492, + "step": 4260 + }, + { + "epoch": 0.619042441375811, + "grad_norm": 1.9921875, + "learning_rate": 1.3811249637576109e-05, + "loss": 0.2652, + "step": 4270 + }, + { + "epoch": 0.6204921894820775, + "grad_norm": 2.5, + "learning_rate": 1.3796752681936793e-05, + "loss": 0.2758, + "step": 4280 + }, + { + "epoch": 0.621941937588344, + "grad_norm": 1.9609375, + "learning_rate": 1.3782255726297478e-05, + "loss": 0.2744, + "step": 4290 + }, + { + "epoch": 0.6233916856946106, + "grad_norm": 2.03125, + "learning_rate": 1.3767758770658164e-05, + "loss": 0.2917, + "step": 4300 + }, + { + "epoch": 0.6248414338008771, + "grad_norm": 2.578125, + "learning_rate": 1.3753261815018848e-05, + "loss": 0.3037, + "step": 4310 + }, + { + "epoch": 0.6262911819071436, + "grad_norm": 1.78125, + "learning_rate": 1.3738764859379533e-05, + "loss": 0.2106, + "step": 4320 + }, + { + "epoch": 0.6277409300134101, + "grad_norm": 2.109375, + "learning_rate": 1.3724267903740215e-05, + "loss": 0.2452, + "step": 4330 + }, + { + "epoch": 0.6291906781196767, + "grad_norm": 2.3125, + "learning_rate": 1.37097709481009e-05, + "loss": 0.3048, + "step": 4340 + }, + { + "epoch": 0.6306404262259433, + "grad_norm": 2.21875, + "learning_rate": 1.3695273992461584e-05, + "loss": 0.2839, + "step": 4350 + }, + { + "epoch": 0.6320901743322098, + "grad_norm": 2.03125, + "learning_rate": 1.3680777036822269e-05, + "loss": 0.2875, + "step": 4360 + }, + { + "epoch": 0.6335399224384763, + "grad_norm": 2.1875, + "learning_rate": 1.3666280081182951e-05, + "loss": 0.2686, + "step": 4370 + }, + { + "epoch": 0.6349896705447429, + "grad_norm": 2.296875, + "learning_rate": 1.3651783125543636e-05, + "loss": 0.279, + "step": 4380 + }, + { + "epoch": 0.6364394186510094, + "grad_norm": 2.15625, + "learning_rate": 1.363728616990432e-05, + "loss": 0.2709, + "step": 4390 + }, + { + "epoch": 0.6378891667572759, + "grad_norm": 1.859375, + "learning_rate": 1.3622789214265006e-05, + "loss": 0.2722, + "step": 4400 + }, + { + "epoch": 0.6393389148635424, + "grad_norm": 2.359375, + "learning_rate": 1.360829225862569e-05, + "loss": 0.2585, + "step": 4410 + }, + { + "epoch": 0.640788662969809, + "grad_norm": 2.3125, + "learning_rate": 1.3593795302986375e-05, + "loss": 0.3041, + "step": 4420 + }, + { + "epoch": 0.6422384110760755, + "grad_norm": 1.953125, + "learning_rate": 1.3579298347347058e-05, + "loss": 0.2763, + "step": 4430 + }, + { + "epoch": 0.6436881591823421, + "grad_norm": 1.8828125, + "learning_rate": 1.3564801391707742e-05, + "loss": 0.2839, + "step": 4440 + }, + { + "epoch": 0.6451379072886086, + "grad_norm": 2.34375, + "learning_rate": 1.3550304436068427e-05, + "loss": 0.2862, + "step": 4450 + }, + { + "epoch": 0.6465876553948752, + "grad_norm": 2.046875, + "learning_rate": 1.3535807480429111e-05, + "loss": 0.2549, + "step": 4460 + }, + { + "epoch": 0.6480374035011417, + "grad_norm": 2.21875, + "learning_rate": 1.3521310524789794e-05, + "loss": 0.3009, + "step": 4470 + }, + { + "epoch": 0.6494871516074082, + "grad_norm": 2.0, + "learning_rate": 1.3506813569150478e-05, + "loss": 0.2688, + "step": 4480 + }, + { + "epoch": 0.6509368997136747, + "grad_norm": 1.78125, + "learning_rate": 1.3492316613511163e-05, + "loss": 0.2365, + "step": 4490 + }, + { + "epoch": 0.6523866478199413, + "grad_norm": 1.8359375, + "learning_rate": 1.3477819657871849e-05, + "loss": 0.2597, + "step": 4500 + }, + { + "epoch": 0.6538363959262078, + "grad_norm": 1.8828125, + "learning_rate": 1.3463322702232533e-05, + "loss": 0.2769, + "step": 4510 + }, + { + "epoch": 0.6552861440324743, + "grad_norm": 1.953125, + "learning_rate": 1.3448825746593218e-05, + "loss": 0.308, + "step": 4520 + }, + { + "epoch": 0.6567358921387408, + "grad_norm": 2.375, + "learning_rate": 1.34343287909539e-05, + "loss": 0.2582, + "step": 4530 + }, + { + "epoch": 0.6581856402450075, + "grad_norm": 1.75, + "learning_rate": 1.3419831835314585e-05, + "loss": 0.2667, + "step": 4540 + }, + { + "epoch": 0.659635388351274, + "grad_norm": 1.859375, + "learning_rate": 1.340533487967527e-05, + "loss": 0.2508, + "step": 4550 + }, + { + "epoch": 0.6610851364575405, + "grad_norm": 1.9296875, + "learning_rate": 1.3390837924035954e-05, + "loss": 0.2845, + "step": 4560 + }, + { + "epoch": 0.662534884563807, + "grad_norm": 1.8671875, + "learning_rate": 1.3376340968396636e-05, + "loss": 0.2645, + "step": 4570 + }, + { + "epoch": 0.6639846326700736, + "grad_norm": 2.078125, + "learning_rate": 1.336184401275732e-05, + "loss": 0.242, + "step": 4580 + }, + { + "epoch": 0.6654343807763401, + "grad_norm": 2.25, + "learning_rate": 1.3347347057118005e-05, + "loss": 0.2298, + "step": 4590 + }, + { + "epoch": 0.6668841288826066, + "grad_norm": 2.125, + "learning_rate": 1.3332850101478691e-05, + "loss": 0.2543, + "step": 4600 + }, + { + "epoch": 0.6683338769888731, + "grad_norm": 1.8984375, + "learning_rate": 1.3318353145839376e-05, + "loss": 0.2471, + "step": 4610 + }, + { + "epoch": 0.6697836250951397, + "grad_norm": 2.0, + "learning_rate": 1.330385619020006e-05, + "loss": 0.2579, + "step": 4620 + }, + { + "epoch": 0.6712333732014063, + "grad_norm": 2.078125, + "learning_rate": 1.3289359234560743e-05, + "loss": 0.2651, + "step": 4630 + }, + { + "epoch": 0.6726831213076728, + "grad_norm": 1.96875, + "learning_rate": 1.3274862278921427e-05, + "loss": 0.2256, + "step": 4640 + }, + { + "epoch": 0.6741328694139394, + "grad_norm": 2.046875, + "learning_rate": 1.3260365323282112e-05, + "loss": 0.2677, + "step": 4650 + }, + { + "epoch": 0.6755826175202059, + "grad_norm": 1.71875, + "learning_rate": 1.3245868367642796e-05, + "loss": 0.2694, + "step": 4660 + }, + { + "epoch": 0.6770323656264724, + "grad_norm": 1.90625, + "learning_rate": 1.323137141200348e-05, + "loss": 0.2627, + "step": 4670 + }, + { + "epoch": 0.6784821137327389, + "grad_norm": 2.046875, + "learning_rate": 1.3216874456364163e-05, + "loss": 0.2708, + "step": 4680 + }, + { + "epoch": 0.6799318618390054, + "grad_norm": 2.046875, + "learning_rate": 1.3202377500724848e-05, + "loss": 0.2499, + "step": 4690 + }, + { + "epoch": 0.681381609945272, + "grad_norm": 1.6640625, + "learning_rate": 1.3187880545085534e-05, + "loss": 0.2368, + "step": 4700 + }, + { + "epoch": 0.6828313580515385, + "grad_norm": 1.8046875, + "learning_rate": 1.3173383589446218e-05, + "loss": 0.2566, + "step": 4710 + }, + { + "epoch": 0.6842811061578051, + "grad_norm": 1.9296875, + "learning_rate": 1.3158886633806903e-05, + "loss": 0.2607, + "step": 4720 + }, + { + "epoch": 0.6857308542640717, + "grad_norm": 1.9375, + "learning_rate": 1.3144389678167585e-05, + "loss": 0.2729, + "step": 4730 + }, + { + "epoch": 0.6871806023703382, + "grad_norm": 1.984375, + "learning_rate": 1.312989272252827e-05, + "loss": 0.2901, + "step": 4740 + }, + { + "epoch": 0.6886303504766047, + "grad_norm": 1.6015625, + "learning_rate": 1.3115395766888954e-05, + "loss": 0.2545, + "step": 4750 + }, + { + "epoch": 0.6900800985828712, + "grad_norm": 2.046875, + "learning_rate": 1.3100898811249639e-05, + "loss": 0.3032, + "step": 4760 + }, + { + "epoch": 0.6915298466891378, + "grad_norm": 1.609375, + "learning_rate": 1.3086401855610323e-05, + "loss": 0.248, + "step": 4770 + }, + { + "epoch": 0.6929795947954043, + "grad_norm": 1.953125, + "learning_rate": 1.3071904899971006e-05, + "loss": 0.2771, + "step": 4780 + }, + { + "epoch": 0.6944293429016708, + "grad_norm": 1.8125, + "learning_rate": 1.305740794433169e-05, + "loss": 0.2461, + "step": 4790 + }, + { + "epoch": 0.6958790910079373, + "grad_norm": 2.046875, + "learning_rate": 1.3042910988692376e-05, + "loss": 0.2745, + "step": 4800 + }, + { + "epoch": 0.697328839114204, + "grad_norm": 1.765625, + "learning_rate": 1.302841403305306e-05, + "loss": 0.2227, + "step": 4810 + }, + { + "epoch": 0.6987785872204705, + "grad_norm": 2.25, + "learning_rate": 1.3013917077413745e-05, + "loss": 0.2337, + "step": 4820 + }, + { + "epoch": 0.700228335326737, + "grad_norm": 2.265625, + "learning_rate": 1.2999420121774428e-05, + "loss": 0.2619, + "step": 4830 + }, + { + "epoch": 0.7016780834330035, + "grad_norm": 1.8359375, + "learning_rate": 1.2984923166135112e-05, + "loss": 0.2547, + "step": 4840 + }, + { + "epoch": 0.70312783153927, + "grad_norm": 2.140625, + "learning_rate": 1.2970426210495797e-05, + "loss": 0.2535, + "step": 4850 + }, + { + "epoch": 0.7045775796455366, + "grad_norm": 1.84375, + "learning_rate": 1.2955929254856481e-05, + "loss": 0.2413, + "step": 4860 + }, + { + "epoch": 0.7060273277518031, + "grad_norm": 1.875, + "learning_rate": 1.2941432299217166e-05, + "loss": 0.24, + "step": 4870 + }, + { + "epoch": 0.7074770758580696, + "grad_norm": 2.203125, + "learning_rate": 1.2926935343577848e-05, + "loss": 0.2506, + "step": 4880 + }, + { + "epoch": 0.7089268239643362, + "grad_norm": 2.203125, + "learning_rate": 1.2912438387938533e-05, + "loss": 0.2338, + "step": 4890 + }, + { + "epoch": 0.7103765720706028, + "grad_norm": 1.6875, + "learning_rate": 1.2897941432299219e-05, + "loss": 0.2733, + "step": 4900 + }, + { + "epoch": 0.7118263201768693, + "grad_norm": 2.046875, + "learning_rate": 1.2883444476659903e-05, + "loss": 0.2348, + "step": 4910 + }, + { + "epoch": 0.7132760682831358, + "grad_norm": 1.7265625, + "learning_rate": 1.2868947521020588e-05, + "loss": 0.2806, + "step": 4920 + }, + { + "epoch": 0.7147258163894024, + "grad_norm": 1.5703125, + "learning_rate": 1.2854450565381272e-05, + "loss": 0.2315, + "step": 4930 + }, + { + "epoch": 0.7161755644956689, + "grad_norm": 1.9921875, + "learning_rate": 1.2839953609741955e-05, + "loss": 0.2482, + "step": 4940 + }, + { + "epoch": 0.7176253126019354, + "grad_norm": 1.671875, + "learning_rate": 1.282545665410264e-05, + "loss": 0.2423, + "step": 4950 + }, + { + "epoch": 0.7190750607082019, + "grad_norm": 2.0625, + "learning_rate": 1.2810959698463324e-05, + "loss": 0.2637, + "step": 4960 + }, + { + "epoch": 0.7205248088144685, + "grad_norm": 1.90625, + "learning_rate": 1.2796462742824008e-05, + "loss": 0.1909, + "step": 4970 + }, + { + "epoch": 0.721974556920735, + "grad_norm": 2.0, + "learning_rate": 1.278196578718469e-05, + "loss": 0.2652, + "step": 4980 + }, + { + "epoch": 0.7234243050270016, + "grad_norm": 1.8828125, + "learning_rate": 1.2767468831545375e-05, + "loss": 0.2419, + "step": 4990 + }, + { + "epoch": 0.7248740531332681, + "grad_norm": 1.921875, + "learning_rate": 1.2752971875906061e-05, + "loss": 0.2425, + "step": 5000 + }, + { + "epoch": 0.7263238012395347, + "grad_norm": 1.78125, + "learning_rate": 1.2738474920266746e-05, + "loss": 0.2363, + "step": 5010 + }, + { + "epoch": 0.7277735493458012, + "grad_norm": 1.796875, + "learning_rate": 1.272397796462743e-05, + "loss": 0.2045, + "step": 5020 + }, + { + "epoch": 0.7292232974520677, + "grad_norm": 1.7734375, + "learning_rate": 1.2709481008988115e-05, + "loss": 0.236, + "step": 5030 + }, + { + "epoch": 0.7306730455583342, + "grad_norm": 1.7265625, + "learning_rate": 1.2694984053348797e-05, + "loss": 0.2325, + "step": 5040 + }, + { + "epoch": 0.7321227936646008, + "grad_norm": 1.4921875, + "learning_rate": 1.2680487097709482e-05, + "loss": 0.2425, + "step": 5050 + }, + { + "epoch": 0.7335725417708673, + "grad_norm": 1.640625, + "learning_rate": 1.2665990142070166e-05, + "loss": 0.2518, + "step": 5060 + }, + { + "epoch": 0.7350222898771338, + "grad_norm": 1.9375, + "learning_rate": 1.265149318643085e-05, + "loss": 0.2871, + "step": 5070 + }, + { + "epoch": 0.7364720379834003, + "grad_norm": 1.7265625, + "learning_rate": 1.2636996230791533e-05, + "loss": 0.2491, + "step": 5080 + }, + { + "epoch": 0.737921786089667, + "grad_norm": 2.03125, + "learning_rate": 1.2622499275152218e-05, + "loss": 0.2626, + "step": 5090 + }, + { + "epoch": 0.7393715341959335, + "grad_norm": 1.9453125, + "learning_rate": 1.2608002319512904e-05, + "loss": 0.28, + "step": 5100 + }, + { + "epoch": 0.7408212823022, + "grad_norm": 2.15625, + "learning_rate": 1.2593505363873588e-05, + "loss": 0.2514, + "step": 5110 + }, + { + "epoch": 0.7422710304084665, + "grad_norm": 2.015625, + "learning_rate": 1.2579008408234273e-05, + "loss": 0.2447, + "step": 5120 + }, + { + "epoch": 0.7437207785147331, + "grad_norm": 1.9375, + "learning_rate": 1.2564511452594957e-05, + "loss": 0.2955, + "step": 5130 + }, + { + "epoch": 0.7451705266209996, + "grad_norm": 1.875, + "learning_rate": 1.255001449695564e-05, + "loss": 0.2018, + "step": 5140 + }, + { + "epoch": 0.7466202747272661, + "grad_norm": 2.265625, + "learning_rate": 1.2535517541316324e-05, + "loss": 0.2646, + "step": 5150 + }, + { + "epoch": 0.7480700228335326, + "grad_norm": 1.8671875, + "learning_rate": 1.2521020585677009e-05, + "loss": 0.2455, + "step": 5160 + }, + { + "epoch": 0.7495197709397992, + "grad_norm": 1.8515625, + "learning_rate": 1.2506523630037693e-05, + "loss": 0.2129, + "step": 5170 + }, + { + "epoch": 0.7509695190460658, + "grad_norm": 1.8828125, + "learning_rate": 1.2492026674398376e-05, + "loss": 0.2461, + "step": 5180 + }, + { + "epoch": 0.7524192671523323, + "grad_norm": 2.03125, + "learning_rate": 1.2477529718759063e-05, + "loss": 0.2581, + "step": 5190 + }, + { + "epoch": 0.7538690152585988, + "grad_norm": 1.7734375, + "learning_rate": 1.2463032763119746e-05, + "loss": 0.232, + "step": 5200 + }, + { + "epoch": 0.7553187633648654, + "grad_norm": 2.4375, + "learning_rate": 1.244853580748043e-05, + "loss": 0.2385, + "step": 5210 + }, + { + "epoch": 0.7567685114711319, + "grad_norm": 1.515625, + "learning_rate": 1.2434038851841115e-05, + "loss": 0.2348, + "step": 5220 + }, + { + "epoch": 0.7582182595773984, + "grad_norm": 1.4765625, + "learning_rate": 1.24195418962018e-05, + "loss": 0.2408, + "step": 5230 + }, + { + "epoch": 0.7596680076836649, + "grad_norm": 2.0, + "learning_rate": 1.2405044940562482e-05, + "loss": 0.2063, + "step": 5240 + }, + { + "epoch": 0.7611177557899315, + "grad_norm": 2.015625, + "learning_rate": 1.2390547984923167e-05, + "loss": 0.2283, + "step": 5250 + }, + { + "epoch": 0.762567503896198, + "grad_norm": 2.09375, + "learning_rate": 1.2376051029283851e-05, + "loss": 0.2586, + "step": 5260 + }, + { + "epoch": 0.7640172520024646, + "grad_norm": 1.6875, + "learning_rate": 1.2361554073644535e-05, + "loss": 0.2276, + "step": 5270 + }, + { + "epoch": 0.7654670001087311, + "grad_norm": 1.9296875, + "learning_rate": 1.234705711800522e-05, + "loss": 0.2136, + "step": 5280 + }, + { + "epoch": 0.7669167482149977, + "grad_norm": 1.96875, + "learning_rate": 1.2332560162365906e-05, + "loss": 0.2388, + "step": 5290 + }, + { + "epoch": 0.7683664963212642, + "grad_norm": 2.015625, + "learning_rate": 1.2318063206726589e-05, + "loss": 0.2481, + "step": 5300 + }, + { + "epoch": 0.7698162444275307, + "grad_norm": 1.8046875, + "learning_rate": 1.2303566251087273e-05, + "loss": 0.2364, + "step": 5310 + }, + { + "epoch": 0.7712659925337972, + "grad_norm": 1.78125, + "learning_rate": 1.2289069295447958e-05, + "loss": 0.2209, + "step": 5320 + }, + { + "epoch": 0.7727157406400638, + "grad_norm": 1.96875, + "learning_rate": 1.2274572339808642e-05, + "loss": 0.2342, + "step": 5330 + }, + { + "epoch": 0.7741654887463303, + "grad_norm": 1.6015625, + "learning_rate": 1.2260075384169325e-05, + "loss": 0.2444, + "step": 5340 + }, + { + "epoch": 0.7756152368525968, + "grad_norm": 1.75, + "learning_rate": 1.2245578428530009e-05, + "loss": 0.2056, + "step": 5350 + }, + { + "epoch": 0.7770649849588634, + "grad_norm": 1.9375, + "learning_rate": 1.2231081472890693e-05, + "loss": 0.2596, + "step": 5360 + }, + { + "epoch": 0.77851473306513, + "grad_norm": 1.390625, + "learning_rate": 1.2216584517251378e-05, + "loss": 0.2151, + "step": 5370 + }, + { + "epoch": 0.7799644811713965, + "grad_norm": 1.7109375, + "learning_rate": 1.2202087561612062e-05, + "loss": 0.2188, + "step": 5380 + }, + { + "epoch": 0.781414229277663, + "grad_norm": 1.8984375, + "learning_rate": 1.2187590605972748e-05, + "loss": 0.2514, + "step": 5390 + }, + { + "epoch": 0.7828639773839295, + "grad_norm": 1.7734375, + "learning_rate": 1.2173093650333431e-05, + "loss": 0.2026, + "step": 5400 + }, + { + "epoch": 0.7843137254901961, + "grad_norm": 1.7734375, + "learning_rate": 1.2158596694694116e-05, + "loss": 0.2261, + "step": 5410 + }, + { + "epoch": 0.7857634735964626, + "grad_norm": 1.796875, + "learning_rate": 1.21440997390548e-05, + "loss": 0.219, + "step": 5420 + }, + { + "epoch": 0.7872132217027291, + "grad_norm": 1.96875, + "learning_rate": 1.2129602783415484e-05, + "loss": 0.2468, + "step": 5430 + }, + { + "epoch": 0.7886629698089956, + "grad_norm": 1.875, + "learning_rate": 1.2115105827776167e-05, + "loss": 0.2486, + "step": 5440 + }, + { + "epoch": 0.7901127179152623, + "grad_norm": 1.8984375, + "learning_rate": 1.2100608872136852e-05, + "loss": 0.2162, + "step": 5450 + }, + { + "epoch": 0.7915624660215288, + "grad_norm": 2.328125, + "learning_rate": 1.2086111916497536e-05, + "loss": 0.2382, + "step": 5460 + }, + { + "epoch": 0.7930122141277953, + "grad_norm": 1.71875, + "learning_rate": 1.207161496085822e-05, + "loss": 0.2424, + "step": 5470 + }, + { + "epoch": 0.7944619622340618, + "grad_norm": 2.328125, + "learning_rate": 1.2057118005218905e-05, + "loss": 0.2417, + "step": 5480 + }, + { + "epoch": 0.7959117103403284, + "grad_norm": 2.515625, + "learning_rate": 1.2042621049579591e-05, + "loss": 0.2332, + "step": 5490 + }, + { + "epoch": 0.7973614584465949, + "grad_norm": 1.9375, + "learning_rate": 1.2028124093940274e-05, + "loss": 0.2265, + "step": 5500 + }, + { + "epoch": 0.7988112065528614, + "grad_norm": 2.015625, + "learning_rate": 1.2013627138300958e-05, + "loss": 0.224, + "step": 5510 + }, + { + "epoch": 0.8002609546591279, + "grad_norm": 1.9375, + "learning_rate": 1.1999130182661642e-05, + "loss": 0.2264, + "step": 5520 + }, + { + "epoch": 0.8017107027653945, + "grad_norm": 1.90625, + "learning_rate": 1.1984633227022327e-05, + "loss": 0.2563, + "step": 5530 + }, + { + "epoch": 0.8031604508716611, + "grad_norm": 1.9453125, + "learning_rate": 1.1970136271383011e-05, + "loss": 0.2239, + "step": 5540 + }, + { + "epoch": 0.8046101989779276, + "grad_norm": 1.625, + "learning_rate": 1.1955639315743694e-05, + "loss": 0.2084, + "step": 5550 + }, + { + "epoch": 0.8060599470841942, + "grad_norm": 1.6640625, + "learning_rate": 1.1941142360104378e-05, + "loss": 0.2408, + "step": 5560 + }, + { + "epoch": 0.8075096951904607, + "grad_norm": 1.671875, + "learning_rate": 1.1926645404465063e-05, + "loss": 0.2034, + "step": 5570 + }, + { + "epoch": 0.8089594432967272, + "grad_norm": 2.140625, + "learning_rate": 1.1912148448825747e-05, + "loss": 0.2504, + "step": 5580 + }, + { + "epoch": 0.8104091914029937, + "grad_norm": 1.765625, + "learning_rate": 1.1897651493186433e-05, + "loss": 0.2253, + "step": 5590 + }, + { + "epoch": 0.8118589395092602, + "grad_norm": 2.703125, + "learning_rate": 1.1883154537547116e-05, + "loss": 0.2498, + "step": 5600 + }, + { + "epoch": 0.8133086876155268, + "grad_norm": 1.6171875, + "learning_rate": 1.18686575819078e-05, + "loss": 0.2267, + "step": 5610 + }, + { + "epoch": 0.8147584357217933, + "grad_norm": 1.5546875, + "learning_rate": 1.1854160626268485e-05, + "loss": 0.2274, + "step": 5620 + }, + { + "epoch": 0.8162081838280599, + "grad_norm": 1.96875, + "learning_rate": 1.183966367062917e-05, + "loss": 0.2231, + "step": 5630 + }, + { + "epoch": 0.8176579319343265, + "grad_norm": 1.921875, + "learning_rate": 1.1825166714989854e-05, + "loss": 0.2096, + "step": 5640 + }, + { + "epoch": 0.819107680040593, + "grad_norm": 1.7265625, + "learning_rate": 1.1810669759350536e-05, + "loss": 0.2704, + "step": 5650 + }, + { + "epoch": 0.8205574281468595, + "grad_norm": 1.6875, + "learning_rate": 1.1796172803711221e-05, + "loss": 0.2066, + "step": 5660 + }, + { + "epoch": 0.822007176253126, + "grad_norm": 1.84375, + "learning_rate": 1.1781675848071905e-05, + "loss": 0.2283, + "step": 5670 + }, + { + "epoch": 0.8234569243593926, + "grad_norm": 1.6875, + "learning_rate": 1.176717889243259e-05, + "loss": 0.2513, + "step": 5680 + }, + { + "epoch": 0.8249066724656591, + "grad_norm": 1.7890625, + "learning_rate": 1.1752681936793276e-05, + "loss": 0.2303, + "step": 5690 + }, + { + "epoch": 0.8263564205719256, + "grad_norm": 1.90625, + "learning_rate": 1.173818498115396e-05, + "loss": 0.2601, + "step": 5700 + }, + { + "epoch": 0.8278061686781921, + "grad_norm": 1.8515625, + "learning_rate": 1.1723688025514643e-05, + "loss": 0.2538, + "step": 5710 + }, + { + "epoch": 0.8292559167844586, + "grad_norm": 1.7734375, + "learning_rate": 1.1709191069875327e-05, + "loss": 0.2514, + "step": 5720 + }, + { + "epoch": 0.8307056648907253, + "grad_norm": 2.046875, + "learning_rate": 1.1694694114236012e-05, + "loss": 0.2452, + "step": 5730 + }, + { + "epoch": 0.8321554129969918, + "grad_norm": 1.7265625, + "learning_rate": 1.1680197158596696e-05, + "loss": 0.2453, + "step": 5740 + }, + { + "epoch": 0.8336051611032583, + "grad_norm": 2.078125, + "learning_rate": 1.1665700202957379e-05, + "loss": 0.2355, + "step": 5750 + }, + { + "epoch": 0.8350549092095249, + "grad_norm": 1.703125, + "learning_rate": 1.1651203247318063e-05, + "loss": 0.2142, + "step": 5760 + }, + { + "epoch": 0.8365046573157914, + "grad_norm": 2.078125, + "learning_rate": 1.1636706291678748e-05, + "loss": 0.2166, + "step": 5770 + }, + { + "epoch": 0.8379544054220579, + "grad_norm": 1.6796875, + "learning_rate": 1.1622209336039432e-05, + "loss": 0.2086, + "step": 5780 + }, + { + "epoch": 0.8394041535283244, + "grad_norm": 1.7578125, + "learning_rate": 1.1607712380400118e-05, + "loss": 0.2362, + "step": 5790 + }, + { + "epoch": 0.840853901634591, + "grad_norm": 1.9609375, + "learning_rate": 1.1593215424760803e-05, + "loss": 0.2538, + "step": 5800 + }, + { + "epoch": 0.8423036497408575, + "grad_norm": 1.8515625, + "learning_rate": 1.1578718469121485e-05, + "loss": 0.2435, + "step": 5810 + }, + { + "epoch": 0.8437533978471241, + "grad_norm": 2.0625, + "learning_rate": 1.156422151348217e-05, + "loss": 0.2312, + "step": 5820 + }, + { + "epoch": 0.8452031459533906, + "grad_norm": 2.125, + "learning_rate": 1.1549724557842854e-05, + "loss": 0.2263, + "step": 5830 + }, + { + "epoch": 0.8466528940596572, + "grad_norm": 1.9453125, + "learning_rate": 1.1535227602203539e-05, + "loss": 0.2159, + "step": 5840 + }, + { + "epoch": 0.8481026421659237, + "grad_norm": 1.578125, + "learning_rate": 1.1520730646564221e-05, + "loss": 0.2498, + "step": 5850 + }, + { + "epoch": 0.8495523902721902, + "grad_norm": 1.75, + "learning_rate": 1.1506233690924906e-05, + "loss": 0.2543, + "step": 5860 + }, + { + "epoch": 0.8510021383784567, + "grad_norm": 1.9140625, + "learning_rate": 1.149173673528559e-05, + "loss": 0.2424, + "step": 5870 + }, + { + "epoch": 0.8524518864847233, + "grad_norm": 2.125, + "learning_rate": 1.1477239779646275e-05, + "loss": 0.2203, + "step": 5880 + }, + { + "epoch": 0.8539016345909898, + "grad_norm": 1.640625, + "learning_rate": 1.146274282400696e-05, + "loss": 0.2512, + "step": 5890 + }, + { + "epoch": 0.8553513826972563, + "grad_norm": 1.734375, + "learning_rate": 1.1448245868367645e-05, + "loss": 0.2151, + "step": 5900 + }, + { + "epoch": 0.8568011308035229, + "grad_norm": 2.171875, + "learning_rate": 1.1433748912728328e-05, + "loss": 0.25, + "step": 5910 + }, + { + "epoch": 0.8582508789097895, + "grad_norm": 1.875, + "learning_rate": 1.1419251957089012e-05, + "loss": 0.2649, + "step": 5920 + }, + { + "epoch": 0.859700627016056, + "grad_norm": 1.6875, + "learning_rate": 1.1404755001449697e-05, + "loss": 0.223, + "step": 5930 + }, + { + "epoch": 0.8611503751223225, + "grad_norm": 1.734375, + "learning_rate": 1.1390258045810381e-05, + "loss": 0.2157, + "step": 5940 + }, + { + "epoch": 0.862600123228589, + "grad_norm": 1.4609375, + "learning_rate": 1.1375761090171064e-05, + "loss": 0.1802, + "step": 5950 + }, + { + "epoch": 0.8640498713348556, + "grad_norm": 1.6484375, + "learning_rate": 1.1361264134531748e-05, + "loss": 0.2185, + "step": 5960 + }, + { + "epoch": 0.8654996194411221, + "grad_norm": 1.640625, + "learning_rate": 1.1346767178892433e-05, + "loss": 0.2055, + "step": 5970 + }, + { + "epoch": 0.8669493675473886, + "grad_norm": 2.1875, + "learning_rate": 1.1332270223253117e-05, + "loss": 0.2157, + "step": 5980 + }, + { + "epoch": 0.8683991156536551, + "grad_norm": 1.9453125, + "learning_rate": 1.1317773267613803e-05, + "loss": 0.2298, + "step": 5990 + }, + { + "epoch": 0.8698488637599218, + "grad_norm": 1.71875, + "learning_rate": 1.1303276311974488e-05, + "loss": 0.2106, + "step": 6000 + }, + { + "epoch": 0.8712986118661883, + "grad_norm": 1.671875, + "learning_rate": 1.128877935633517e-05, + "loss": 0.2093, + "step": 6010 + }, + { + "epoch": 0.8727483599724548, + "grad_norm": 1.90625, + "learning_rate": 1.1274282400695855e-05, + "loss": 0.2418, + "step": 6020 + }, + { + "epoch": 0.8741981080787213, + "grad_norm": 2.125, + "learning_rate": 1.125978544505654e-05, + "loss": 0.2467, + "step": 6030 + }, + { + "epoch": 0.8756478561849879, + "grad_norm": 2.03125, + "learning_rate": 1.1245288489417224e-05, + "loss": 0.2458, + "step": 6040 + }, + { + "epoch": 0.8770976042912544, + "grad_norm": 1.8046875, + "learning_rate": 1.1230791533777906e-05, + "loss": 0.2425, + "step": 6050 + }, + { + "epoch": 0.8785473523975209, + "grad_norm": 1.5234375, + "learning_rate": 1.121629457813859e-05, + "loss": 0.2363, + "step": 6060 + }, + { + "epoch": 0.8799971005037874, + "grad_norm": 1.796875, + "learning_rate": 1.1201797622499275e-05, + "loss": 0.2397, + "step": 6070 + }, + { + "epoch": 0.881446848610054, + "grad_norm": 1.5390625, + "learning_rate": 1.118730066685996e-05, + "loss": 0.2139, + "step": 6080 + }, + { + "epoch": 0.8828965967163206, + "grad_norm": 1.6796875, + "learning_rate": 1.1172803711220646e-05, + "loss": 0.229, + "step": 6090 + }, + { + "epoch": 0.8843463448225871, + "grad_norm": 1.7109375, + "learning_rate": 1.115830675558133e-05, + "loss": 0.2004, + "step": 6100 + }, + { + "epoch": 0.8857960929288536, + "grad_norm": 1.890625, + "learning_rate": 1.1143809799942013e-05, + "loss": 0.2233, + "step": 6110 + }, + { + "epoch": 0.8872458410351202, + "grad_norm": 1.8359375, + "learning_rate": 1.1129312844302697e-05, + "loss": 0.2162, + "step": 6120 + }, + { + "epoch": 0.8886955891413867, + "grad_norm": 1.59375, + "learning_rate": 1.1114815888663382e-05, + "loss": 0.2619, + "step": 6130 + }, + { + "epoch": 0.8901453372476532, + "grad_norm": 1.84375, + "learning_rate": 1.1100318933024066e-05, + "loss": 0.2318, + "step": 6140 + }, + { + "epoch": 0.8915950853539197, + "grad_norm": 2.15625, + "learning_rate": 1.108582197738475e-05, + "loss": 0.2036, + "step": 6150 + }, + { + "epoch": 0.8930448334601863, + "grad_norm": 1.4765625, + "learning_rate": 1.1071325021745433e-05, + "loss": 0.2361, + "step": 6160 + }, + { + "epoch": 0.8944945815664528, + "grad_norm": 1.859375, + "learning_rate": 1.1056828066106118e-05, + "loss": 0.2239, + "step": 6170 + }, + { + "epoch": 0.8959443296727194, + "grad_norm": 1.75, + "learning_rate": 1.1042331110466802e-05, + "loss": 0.2337, + "step": 6180 + }, + { + "epoch": 0.8973940777789859, + "grad_norm": 2.21875, + "learning_rate": 1.1027834154827488e-05, + "loss": 0.23, + "step": 6190 + }, + { + "epoch": 0.8988438258852525, + "grad_norm": 1.734375, + "learning_rate": 1.1013337199188173e-05, + "loss": 0.2014, + "step": 6200 + }, + { + "epoch": 0.900293573991519, + "grad_norm": 1.5859375, + "learning_rate": 1.0998840243548855e-05, + "loss": 0.2323, + "step": 6210 + }, + { + "epoch": 0.9017433220977855, + "grad_norm": 1.296875, + "learning_rate": 1.098434328790954e-05, + "loss": 0.1748, + "step": 6220 + }, + { + "epoch": 0.903193070204052, + "grad_norm": 2.09375, + "learning_rate": 1.0969846332270224e-05, + "loss": 0.2295, + "step": 6230 + }, + { + "epoch": 0.9046428183103186, + "grad_norm": 1.40625, + "learning_rate": 1.0955349376630909e-05, + "loss": 0.2217, + "step": 6240 + }, + { + "epoch": 0.9060925664165851, + "grad_norm": 1.890625, + "learning_rate": 1.0940852420991593e-05, + "loss": 0.197, + "step": 6250 + }, + { + "epoch": 0.9075423145228516, + "grad_norm": 2.0, + "learning_rate": 1.0926355465352276e-05, + "loss": 0.2455, + "step": 6260 + }, + { + "epoch": 0.9089920626291181, + "grad_norm": 1.578125, + "learning_rate": 1.091185850971296e-05, + "loss": 0.1941, + "step": 6270 + }, + { + "epoch": 0.9104418107353848, + "grad_norm": 1.9296875, + "learning_rate": 1.0897361554073645e-05, + "loss": 0.2331, + "step": 6280 + }, + { + "epoch": 0.9118915588416513, + "grad_norm": 1.59375, + "learning_rate": 1.088286459843433e-05, + "loss": 0.2084, + "step": 6290 + }, + { + "epoch": 0.9133413069479178, + "grad_norm": 1.7421875, + "learning_rate": 1.0868367642795015e-05, + "loss": 0.2208, + "step": 6300 + }, + { + "epoch": 0.9147910550541843, + "grad_norm": 2.265625, + "learning_rate": 1.08538706871557e-05, + "loss": 0.2248, + "step": 6310 + }, + { + "epoch": 0.9162408031604509, + "grad_norm": 1.7421875, + "learning_rate": 1.0839373731516382e-05, + "loss": 0.2357, + "step": 6320 + }, + { + "epoch": 0.9176905512667174, + "grad_norm": 1.375, + "learning_rate": 1.0824876775877067e-05, + "loss": 0.2225, + "step": 6330 + }, + { + "epoch": 0.9191402993729839, + "grad_norm": 1.7578125, + "learning_rate": 1.0810379820237751e-05, + "loss": 0.1926, + "step": 6340 + }, + { + "epoch": 0.9205900474792504, + "grad_norm": 2.0625, + "learning_rate": 1.0795882864598435e-05, + "loss": 0.1965, + "step": 6350 + }, + { + "epoch": 0.922039795585517, + "grad_norm": 2.203125, + "learning_rate": 1.0781385908959118e-05, + "loss": 0.2076, + "step": 6360 + }, + { + "epoch": 0.9234895436917836, + "grad_norm": 1.65625, + "learning_rate": 1.0766888953319803e-05, + "loss": 0.2054, + "step": 6370 + }, + { + "epoch": 0.9249392917980501, + "grad_norm": 1.515625, + "learning_rate": 1.0752391997680487e-05, + "loss": 0.2021, + "step": 6380 + }, + { + "epoch": 0.9263890399043166, + "grad_norm": 1.484375, + "learning_rate": 1.0737895042041173e-05, + "loss": 0.1864, + "step": 6390 + }, + { + "epoch": 0.9278387880105832, + "grad_norm": 2.28125, + "learning_rate": 1.0723398086401858e-05, + "loss": 0.1947, + "step": 6400 + }, + { + "epoch": 0.9292885361168497, + "grad_norm": 1.7109375, + "learning_rate": 1.0708901130762542e-05, + "loss": 0.2233, + "step": 6410 + }, + { + "epoch": 0.9307382842231162, + "grad_norm": 1.671875, + "learning_rate": 1.0694404175123225e-05, + "loss": 0.2042, + "step": 6420 + }, + { + "epoch": 0.9321880323293827, + "grad_norm": 1.9609375, + "learning_rate": 1.0679907219483909e-05, + "loss": 0.1938, + "step": 6430 + }, + { + "epoch": 0.9336377804356493, + "grad_norm": 1.8984375, + "learning_rate": 1.0665410263844594e-05, + "loss": 0.2292, + "step": 6440 + }, + { + "epoch": 0.9350875285419158, + "grad_norm": 1.859375, + "learning_rate": 1.0650913308205278e-05, + "loss": 0.188, + "step": 6450 + }, + { + "epoch": 0.9365372766481824, + "grad_norm": 1.859375, + "learning_rate": 1.063641635256596e-05, + "loss": 0.2111, + "step": 6460 + }, + { + "epoch": 0.937987024754449, + "grad_norm": 1.4296875, + "learning_rate": 1.0621919396926645e-05, + "loss": 0.1935, + "step": 6470 + }, + { + "epoch": 0.9394367728607155, + "grad_norm": 1.65625, + "learning_rate": 1.060742244128733e-05, + "loss": 0.2337, + "step": 6480 + }, + { + "epoch": 0.940886520966982, + "grad_norm": 1.6171875, + "learning_rate": 1.0592925485648016e-05, + "loss": 0.217, + "step": 6490 + }, + { + "epoch": 0.9423362690732485, + "grad_norm": 1.640625, + "learning_rate": 1.05784285300087e-05, + "loss": 0.1999, + "step": 6500 + }, + { + "epoch": 0.943786017179515, + "grad_norm": 1.3984375, + "learning_rate": 1.0563931574369384e-05, + "loss": 0.2368, + "step": 6510 + }, + { + "epoch": 0.9452357652857816, + "grad_norm": 1.8515625, + "learning_rate": 1.0549434618730067e-05, + "loss": 0.2015, + "step": 6520 + }, + { + "epoch": 0.9466855133920481, + "grad_norm": 1.28125, + "learning_rate": 1.0534937663090752e-05, + "loss": 0.1905, + "step": 6530 + }, + { + "epoch": 0.9481352614983146, + "grad_norm": 1.6640625, + "learning_rate": 1.0520440707451436e-05, + "loss": 0.211, + "step": 6540 + }, + { + "epoch": 0.9495850096045813, + "grad_norm": 1.7421875, + "learning_rate": 1.050594375181212e-05, + "loss": 0.218, + "step": 6550 + }, + { + "epoch": 0.9510347577108478, + "grad_norm": 1.84375, + "learning_rate": 1.0491446796172803e-05, + "loss": 0.2278, + "step": 6560 + }, + { + "epoch": 0.9524845058171143, + "grad_norm": 1.5625, + "learning_rate": 1.0476949840533488e-05, + "loss": 0.2265, + "step": 6570 + }, + { + "epoch": 0.9539342539233808, + "grad_norm": 1.65625, + "learning_rate": 1.0462452884894172e-05, + "loss": 0.2122, + "step": 6580 + }, + { + "epoch": 0.9553840020296474, + "grad_norm": 2.203125, + "learning_rate": 1.0447955929254858e-05, + "loss": 0.1977, + "step": 6590 + }, + { + "epoch": 0.9568337501359139, + "grad_norm": 1.703125, + "learning_rate": 1.0433458973615543e-05, + "loss": 0.2086, + "step": 6600 + }, + { + "epoch": 0.9582834982421804, + "grad_norm": 1.46875, + "learning_rate": 1.0418962017976227e-05, + "loss": 0.2153, + "step": 6610 + }, + { + "epoch": 0.9597332463484469, + "grad_norm": 1.703125, + "learning_rate": 1.040446506233691e-05, + "loss": 0.1882, + "step": 6620 + }, + { + "epoch": 0.9611829944547134, + "grad_norm": 1.921875, + "learning_rate": 1.0389968106697594e-05, + "loss": 0.2172, + "step": 6630 + }, + { + "epoch": 0.9626327425609801, + "grad_norm": 1.453125, + "learning_rate": 1.0375471151058278e-05, + "loss": 0.1949, + "step": 6640 + }, + { + "epoch": 0.9640824906672466, + "grad_norm": 1.40625, + "learning_rate": 1.0360974195418963e-05, + "loss": 0.2007, + "step": 6650 + }, + { + "epoch": 0.9655322387735131, + "grad_norm": 2.203125, + "learning_rate": 1.0346477239779646e-05, + "loss": 0.2231, + "step": 6660 + }, + { + "epoch": 0.9669819868797797, + "grad_norm": 1.234375, + "learning_rate": 1.033198028414033e-05, + "loss": 0.1793, + "step": 6670 + }, + { + "epoch": 0.9684317349860462, + "grad_norm": 1.703125, + "learning_rate": 1.0317483328501014e-05, + "loss": 0.2123, + "step": 6680 + }, + { + "epoch": 0.9698814830923127, + "grad_norm": 1.734375, + "learning_rate": 1.03029863728617e-05, + "loss": 0.2037, + "step": 6690 + }, + { + "epoch": 0.9713312311985792, + "grad_norm": 1.796875, + "learning_rate": 1.0288489417222385e-05, + "loss": 0.1833, + "step": 6700 + }, + { + "epoch": 0.9727809793048457, + "grad_norm": 2.21875, + "learning_rate": 1.027399246158307e-05, + "loss": 0.2333, + "step": 6710 + }, + { + "epoch": 0.9742307274111123, + "grad_norm": 1.6875, + "learning_rate": 1.0259495505943752e-05, + "loss": 0.2051, + "step": 6720 + }, + { + "epoch": 0.9756804755173789, + "grad_norm": 1.671875, + "learning_rate": 1.0244998550304437e-05, + "loss": 0.2282, + "step": 6730 + }, + { + "epoch": 0.9771302236236454, + "grad_norm": 1.734375, + "learning_rate": 1.0230501594665121e-05, + "loss": 0.2104, + "step": 6740 + }, + { + "epoch": 0.978579971729912, + "grad_norm": 1.8828125, + "learning_rate": 1.0216004639025805e-05, + "loss": 0.2206, + "step": 6750 + }, + { + "epoch": 0.9800297198361785, + "grad_norm": 2.109375, + "learning_rate": 1.020150768338649e-05, + "loss": 0.1934, + "step": 6760 + }, + { + "epoch": 0.981479467942445, + "grad_norm": 2.046875, + "learning_rate": 1.0187010727747173e-05, + "loss": 0.2156, + "step": 6770 + }, + { + "epoch": 0.9829292160487115, + "grad_norm": 1.78125, + "learning_rate": 1.0172513772107859e-05, + "loss": 0.1732, + "step": 6780 + }, + { + "epoch": 0.984378964154978, + "grad_norm": 1.703125, + "learning_rate": 1.0158016816468543e-05, + "loss": 0.2188, + "step": 6790 + }, + { + "epoch": 0.9858287122612446, + "grad_norm": 1.7109375, + "learning_rate": 1.0143519860829227e-05, + "loss": 0.2172, + "step": 6800 + }, + { + "epoch": 0.9872784603675111, + "grad_norm": 1.765625, + "learning_rate": 1.0129022905189912e-05, + "loss": 0.2548, + "step": 6810 + }, + { + "epoch": 0.9887282084737777, + "grad_norm": 1.921875, + "learning_rate": 1.0114525949550595e-05, + "loss": 0.2045, + "step": 6820 + }, + { + "epoch": 0.9901779565800443, + "grad_norm": 1.7890625, + "learning_rate": 1.0100028993911279e-05, + "loss": 0.1958, + "step": 6830 + }, + { + "epoch": 0.9916277046863108, + "grad_norm": 1.859375, + "learning_rate": 1.0085532038271963e-05, + "loss": 0.2075, + "step": 6840 + }, + { + "epoch": 0.9930774527925773, + "grad_norm": 1.6015625, + "learning_rate": 1.0071035082632648e-05, + "loss": 0.1962, + "step": 6850 + }, + { + "epoch": 0.9945272008988438, + "grad_norm": 2.234375, + "learning_rate": 1.0056538126993332e-05, + "loss": 0.2256, + "step": 6860 + }, + { + "epoch": 0.9959769490051104, + "grad_norm": 1.8203125, + "learning_rate": 1.0042041171354015e-05, + "loss": 0.2345, + "step": 6870 + }, + { + "epoch": 0.9974266971113769, + "grad_norm": 1.65625, + "learning_rate": 1.0027544215714701e-05, + "loss": 0.2048, + "step": 6880 + }, + { + "epoch": 0.9988764452176434, + "grad_norm": 1.75, + "learning_rate": 1.0013047260075386e-05, + "loss": 0.1975, + "step": 6890 + }, + { + "epoch": 1.0002899496212534, + "grad_norm": 1.53125, + "learning_rate": 9.998550304436068e-06, + "loss": 0.1921, + "step": 6900 + }, + { + "epoch": 1.00173969772752, + "grad_norm": 1.4609375, + "learning_rate": 9.984053348796754e-06, + "loss": 0.1598, + "step": 6910 + }, + { + "epoch": 1.0031894458337864, + "grad_norm": 1.7734375, + "learning_rate": 9.969556393157439e-06, + "loss": 0.1326, + "step": 6920 + }, + { + "epoch": 1.004639193940053, + "grad_norm": 1.6015625, + "learning_rate": 9.955059437518121e-06, + "loss": 0.1603, + "step": 6930 + }, + { + "epoch": 1.0060889420463195, + "grad_norm": 1.5703125, + "learning_rate": 9.940562481878806e-06, + "loss": 0.1416, + "step": 6940 + }, + { + "epoch": 1.007538690152586, + "grad_norm": 1.59375, + "learning_rate": 9.92606552623949e-06, + "loss": 0.1626, + "step": 6950 + }, + { + "epoch": 1.0089884382588525, + "grad_norm": 2.03125, + "learning_rate": 9.911568570600175e-06, + "loss": 0.1731, + "step": 6960 + }, + { + "epoch": 1.010438186365119, + "grad_norm": 1.390625, + "learning_rate": 9.89707161496086e-06, + "loss": 0.1496, + "step": 6970 + }, + { + "epoch": 1.0118879344713856, + "grad_norm": 1.8046875, + "learning_rate": 9.882574659321544e-06, + "loss": 0.2041, + "step": 6980 + }, + { + "epoch": 1.013337682577652, + "grad_norm": 1.6171875, + "learning_rate": 9.868077703682228e-06, + "loss": 0.1567, + "step": 6990 + }, + { + "epoch": 1.0147874306839186, + "grad_norm": 1.171875, + "learning_rate": 9.85358074804291e-06, + "loss": 0.1432, + "step": 7000 + }, + { + "epoch": 1.0162371787901852, + "grad_norm": 1.5546875, + "learning_rate": 9.839083792403597e-06, + "loss": 0.1426, + "step": 7010 + }, + { + "epoch": 1.0176869268964517, + "grad_norm": 1.625, + "learning_rate": 9.824586836764281e-06, + "loss": 0.1792, + "step": 7020 + }, + { + "epoch": 1.0191366750027182, + "grad_norm": 1.421875, + "learning_rate": 9.810089881124964e-06, + "loss": 0.1704, + "step": 7030 + }, + { + "epoch": 1.0205864231089847, + "grad_norm": 1.28125, + "learning_rate": 9.795592925485648e-06, + "loss": 0.1448, + "step": 7040 + }, + { + "epoch": 1.0220361712152513, + "grad_norm": 1.7890625, + "learning_rate": 9.781095969846333e-06, + "loss": 0.1665, + "step": 7050 + }, + { + "epoch": 1.0234859193215178, + "grad_norm": 1.46875, + "learning_rate": 9.766599014207017e-06, + "loss": 0.1611, + "step": 7060 + }, + { + "epoch": 1.0249356674277845, + "grad_norm": 1.828125, + "learning_rate": 9.752102058567702e-06, + "loss": 0.147, + "step": 7070 + }, + { + "epoch": 1.026385415534051, + "grad_norm": 1.5546875, + "learning_rate": 9.737605102928386e-06, + "loss": 0.1515, + "step": 7080 + }, + { + "epoch": 1.0278351636403176, + "grad_norm": 1.6015625, + "learning_rate": 9.72310814728907e-06, + "loss": 0.1426, + "step": 7090 + }, + { + "epoch": 1.029284911746584, + "grad_norm": 1.9296875, + "learning_rate": 9.708611191649753e-06, + "loss": 0.1541, + "step": 7100 + }, + { + "epoch": 1.0307346598528506, + "grad_norm": 1.625, + "learning_rate": 9.69411423601044e-06, + "loss": 0.1467, + "step": 7110 + }, + { + "epoch": 1.0321844079591171, + "grad_norm": 1.8671875, + "learning_rate": 9.679617280371124e-06, + "loss": 0.1584, + "step": 7120 + }, + { + "epoch": 1.0336341560653837, + "grad_norm": 1.6484375, + "learning_rate": 9.665120324731806e-06, + "loss": 0.135, + "step": 7130 + }, + { + "epoch": 1.0350839041716502, + "grad_norm": 1.5234375, + "learning_rate": 9.650623369092491e-06, + "loss": 0.1903, + "step": 7140 + }, + { + "epoch": 1.0365336522779167, + "grad_norm": 1.4921875, + "learning_rate": 9.636126413453175e-06, + "loss": 0.1558, + "step": 7150 + }, + { + "epoch": 1.0379834003841832, + "grad_norm": 1.6015625, + "learning_rate": 9.62162945781386e-06, + "loss": 0.1407, + "step": 7160 + }, + { + "epoch": 1.0394331484904498, + "grad_norm": 1.9375, + "learning_rate": 9.607132502174544e-06, + "loss": 0.1712, + "step": 7170 + }, + { + "epoch": 1.0408828965967163, + "grad_norm": 1.46875, + "learning_rate": 9.592635546535229e-06, + "loss": 0.152, + "step": 7180 + }, + { + "epoch": 1.0423326447029828, + "grad_norm": 1.78125, + "learning_rate": 9.578138590895913e-06, + "loss": 0.1636, + "step": 7190 + }, + { + "epoch": 1.0437823928092493, + "grad_norm": 1.6953125, + "learning_rate": 9.563641635256596e-06, + "loss": 0.1673, + "step": 7200 + }, + { + "epoch": 1.0452321409155159, + "grad_norm": 1.8125, + "learning_rate": 9.549144679617282e-06, + "loss": 0.155, + "step": 7210 + }, + { + "epoch": 1.0466818890217824, + "grad_norm": 1.75, + "learning_rate": 9.534647723977966e-06, + "loss": 0.1487, + "step": 7220 + }, + { + "epoch": 1.048131637128049, + "grad_norm": 1.5078125, + "learning_rate": 9.520150768338649e-06, + "loss": 0.1465, + "step": 7230 + }, + { + "epoch": 1.0495813852343154, + "grad_norm": 1.734375, + "learning_rate": 9.505653812699333e-06, + "loss": 0.174, + "step": 7240 + }, + { + "epoch": 1.0510311333405822, + "grad_norm": 1.7109375, + "learning_rate": 9.491156857060018e-06, + "loss": 0.1594, + "step": 7250 + }, + { + "epoch": 1.0524808814468487, + "grad_norm": 1.5546875, + "learning_rate": 9.476659901420702e-06, + "loss": 0.1531, + "step": 7260 + }, + { + "epoch": 1.0539306295531152, + "grad_norm": 1.5859375, + "learning_rate": 9.462162945781387e-06, + "loss": 0.1459, + "step": 7270 + }, + { + "epoch": 1.0553803776593818, + "grad_norm": 1.09375, + "learning_rate": 9.447665990142071e-06, + "loss": 0.1616, + "step": 7280 + }, + { + "epoch": 1.0568301257656483, + "grad_norm": 1.296875, + "learning_rate": 9.433169034502755e-06, + "loss": 0.1244, + "step": 7290 + }, + { + "epoch": 1.0582798738719148, + "grad_norm": 1.5234375, + "learning_rate": 9.418672078863438e-06, + "loss": 0.1475, + "step": 7300 + }, + { + "epoch": 1.0597296219781813, + "grad_norm": 1.5703125, + "learning_rate": 9.404175123224124e-06, + "loss": 0.1433, + "step": 7310 + }, + { + "epoch": 1.0611793700844478, + "grad_norm": 1.421875, + "learning_rate": 9.389678167584809e-06, + "loss": 0.1458, + "step": 7320 + }, + { + "epoch": 1.0626291181907144, + "grad_norm": 1.7421875, + "learning_rate": 9.375181211945491e-06, + "loss": 0.1605, + "step": 7330 + }, + { + "epoch": 1.064078866296981, + "grad_norm": 1.8984375, + "learning_rate": 9.360684256306176e-06, + "loss": 0.1525, + "step": 7340 + }, + { + "epoch": 1.0655286144032474, + "grad_norm": 1.734375, + "learning_rate": 9.34618730066686e-06, + "loss": 0.1348, + "step": 7350 + }, + { + "epoch": 1.066978362509514, + "grad_norm": 1.4453125, + "learning_rate": 9.331690345027545e-06, + "loss": 0.145, + "step": 7360 + }, + { + "epoch": 1.0684281106157805, + "grad_norm": 1.734375, + "learning_rate": 9.317193389388229e-06, + "loss": 0.15, + "step": 7370 + }, + { + "epoch": 1.069877858722047, + "grad_norm": 1.5546875, + "learning_rate": 9.302696433748913e-06, + "loss": 0.1474, + "step": 7380 + }, + { + "epoch": 1.0713276068283135, + "grad_norm": 1.9296875, + "learning_rate": 9.288199478109598e-06, + "loss": 0.1575, + "step": 7390 + }, + { + "epoch": 1.07277735493458, + "grad_norm": 1.4375, + "learning_rate": 9.273702522470282e-06, + "loss": 0.1481, + "step": 7400 + }, + { + "epoch": 1.0742271030408466, + "grad_norm": 2.171875, + "learning_rate": 9.259205566830967e-06, + "loss": 0.1754, + "step": 7410 + }, + { + "epoch": 1.075676851147113, + "grad_norm": 1.53125, + "learning_rate": 9.244708611191651e-06, + "loss": 0.1724, + "step": 7420 + }, + { + "epoch": 1.0771265992533796, + "grad_norm": 1.3671875, + "learning_rate": 9.230211655552334e-06, + "loss": 0.1561, + "step": 7430 + }, + { + "epoch": 1.0785763473596464, + "grad_norm": 1.640625, + "learning_rate": 9.215714699913018e-06, + "loss": 0.1524, + "step": 7440 + }, + { + "epoch": 1.0800260954659129, + "grad_norm": 1.609375, + "learning_rate": 9.201217744273703e-06, + "loss": 0.1552, + "step": 7450 + }, + { + "epoch": 1.0814758435721794, + "grad_norm": 1.7578125, + "learning_rate": 9.186720788634387e-06, + "loss": 0.1723, + "step": 7460 + }, + { + "epoch": 1.082925591678446, + "grad_norm": 1.9453125, + "learning_rate": 9.172223832995072e-06, + "loss": 0.159, + "step": 7470 + }, + { + "epoch": 1.0843753397847125, + "grad_norm": 1.5390625, + "learning_rate": 9.157726877355756e-06, + "loss": 0.1582, + "step": 7480 + }, + { + "epoch": 1.085825087890979, + "grad_norm": 1.734375, + "learning_rate": 9.14322992171644e-06, + "loss": 0.1525, + "step": 7490 + }, + { + "epoch": 1.0872748359972455, + "grad_norm": 1.6640625, + "learning_rate": 9.128732966077125e-06, + "loss": 0.1524, + "step": 7500 + }, + { + "epoch": 1.088724584103512, + "grad_norm": 1.53125, + "learning_rate": 9.11423601043781e-06, + "loss": 0.1465, + "step": 7510 + }, + { + "epoch": 1.0901743322097786, + "grad_norm": 1.6953125, + "learning_rate": 9.099739054798494e-06, + "loss": 0.1521, + "step": 7520 + }, + { + "epoch": 1.091624080316045, + "grad_norm": 1.90625, + "learning_rate": 9.085242099159178e-06, + "loss": 0.1492, + "step": 7530 + }, + { + "epoch": 1.0930738284223116, + "grad_norm": 1.9140625, + "learning_rate": 9.07074514351986e-06, + "loss": 0.1351, + "step": 7540 + }, + { + "epoch": 1.0945235765285781, + "grad_norm": 1.5390625, + "learning_rate": 9.056248187880545e-06, + "loss": 0.1365, + "step": 7550 + }, + { + "epoch": 1.0959733246348446, + "grad_norm": 1.421875, + "learning_rate": 9.04175123224123e-06, + "loss": 0.1801, + "step": 7560 + }, + { + "epoch": 1.0974230727411112, + "grad_norm": 1.7421875, + "learning_rate": 9.027254276601914e-06, + "loss": 0.1233, + "step": 7570 + }, + { + "epoch": 1.0988728208473777, + "grad_norm": 1.4375, + "learning_rate": 9.012757320962598e-06, + "loss": 0.1306, + "step": 7580 + }, + { + "epoch": 1.1003225689536442, + "grad_norm": 1.6171875, + "learning_rate": 8.998260365323283e-06, + "loss": 0.1463, + "step": 7590 + }, + { + "epoch": 1.1017723170599107, + "grad_norm": 1.8828125, + "learning_rate": 8.983763409683967e-06, + "loss": 0.1423, + "step": 7600 + }, + { + "epoch": 1.1032220651661775, + "grad_norm": 1.5390625, + "learning_rate": 8.969266454044652e-06, + "loss": 0.1553, + "step": 7610 + }, + { + "epoch": 1.104671813272444, + "grad_norm": 1.5703125, + "learning_rate": 8.954769498405336e-06, + "loss": 0.1635, + "step": 7620 + }, + { + "epoch": 1.1061215613787105, + "grad_norm": 1.3671875, + "learning_rate": 8.94027254276602e-06, + "loss": 0.1437, + "step": 7630 + }, + { + "epoch": 1.107571309484977, + "grad_norm": 2.21875, + "learning_rate": 8.925775587126703e-06, + "loss": 0.1281, + "step": 7640 + }, + { + "epoch": 1.1090210575912436, + "grad_norm": 1.6328125, + "learning_rate": 8.91127863148739e-06, + "loss": 0.1542, + "step": 7650 + }, + { + "epoch": 1.11047080569751, + "grad_norm": 2.0625, + "learning_rate": 8.896781675848074e-06, + "loss": 0.1591, + "step": 7660 + }, + { + "epoch": 1.1119205538037766, + "grad_norm": 1.40625, + "learning_rate": 8.882284720208756e-06, + "loss": 0.1549, + "step": 7670 + }, + { + "epoch": 1.1133703019100432, + "grad_norm": 1.5546875, + "learning_rate": 8.867787764569441e-06, + "loss": 0.1521, + "step": 7680 + }, + { + "epoch": 1.1148200500163097, + "grad_norm": 2.25, + "learning_rate": 8.853290808930125e-06, + "loss": 0.1677, + "step": 7690 + }, + { + "epoch": 1.1162697981225762, + "grad_norm": 1.5859375, + "learning_rate": 8.83879385329081e-06, + "loss": 0.156, + "step": 7700 + }, + { + "epoch": 1.1177195462288427, + "grad_norm": 2.015625, + "learning_rate": 8.824296897651494e-06, + "loss": 0.1774, + "step": 7710 + }, + { + "epoch": 1.1191692943351093, + "grad_norm": 1.859375, + "learning_rate": 8.809799942012179e-06, + "loss": 0.146, + "step": 7720 + }, + { + "epoch": 1.1206190424413758, + "grad_norm": 2.109375, + "learning_rate": 8.795302986372863e-06, + "loss": 0.1492, + "step": 7730 + }, + { + "epoch": 1.1220687905476423, + "grad_norm": 1.7265625, + "learning_rate": 8.780806030733546e-06, + "loss": 0.1647, + "step": 7740 + }, + { + "epoch": 1.1235185386539088, + "grad_norm": 1.40625, + "learning_rate": 8.766309075094232e-06, + "loss": 0.1695, + "step": 7750 + }, + { + "epoch": 1.1249682867601754, + "grad_norm": 2.140625, + "learning_rate": 8.751812119454916e-06, + "loss": 0.152, + "step": 7760 + }, + { + "epoch": 1.1264180348664419, + "grad_norm": 1.8359375, + "learning_rate": 8.737315163815599e-06, + "loss": 0.1476, + "step": 7770 + }, + { + "epoch": 1.1278677829727084, + "grad_norm": 1.3125, + "learning_rate": 8.722818208176283e-06, + "loss": 0.1764, + "step": 7780 + }, + { + "epoch": 1.129317531078975, + "grad_norm": 1.578125, + "learning_rate": 8.708321252536968e-06, + "loss": 0.1698, + "step": 7790 + }, + { + "epoch": 1.1307672791852417, + "grad_norm": 2.15625, + "learning_rate": 8.693824296897652e-06, + "loss": 0.1433, + "step": 7800 + }, + { + "epoch": 1.1322170272915082, + "grad_norm": 2.1875, + "learning_rate": 8.679327341258337e-06, + "loss": 0.1604, + "step": 7810 + }, + { + "epoch": 1.1336667753977747, + "grad_norm": 1.4921875, + "learning_rate": 8.664830385619021e-06, + "loss": 0.1373, + "step": 7820 + }, + { + "epoch": 1.1351165235040412, + "grad_norm": 1.0234375, + "learning_rate": 8.650333429979705e-06, + "loss": 0.1478, + "step": 7830 + }, + { + "epoch": 1.1365662716103078, + "grad_norm": 1.4609375, + "learning_rate": 8.635836474340388e-06, + "loss": 0.1596, + "step": 7840 + }, + { + "epoch": 1.1380160197165743, + "grad_norm": 1.65625, + "learning_rate": 8.621339518701074e-06, + "loss": 0.142, + "step": 7850 + }, + { + "epoch": 1.1394657678228408, + "grad_norm": 1.8515625, + "learning_rate": 8.606842563061759e-06, + "loss": 0.1584, + "step": 7860 + }, + { + "epoch": 1.1409155159291073, + "grad_norm": 1.4609375, + "learning_rate": 8.592345607422441e-06, + "loss": 0.1359, + "step": 7870 + }, + { + "epoch": 1.1423652640353739, + "grad_norm": 1.15625, + "learning_rate": 8.577848651783126e-06, + "loss": 0.1619, + "step": 7880 + }, + { + "epoch": 1.1438150121416404, + "grad_norm": 1.5234375, + "learning_rate": 8.56335169614381e-06, + "loss": 0.167, + "step": 7890 + }, + { + "epoch": 1.145264760247907, + "grad_norm": 1.8359375, + "learning_rate": 8.548854740504495e-06, + "loss": 0.1569, + "step": 7900 + }, + { + "epoch": 1.1467145083541734, + "grad_norm": 1.78125, + "learning_rate": 8.534357784865179e-06, + "loss": 0.1508, + "step": 7910 + }, + { + "epoch": 1.14816425646044, + "grad_norm": 1.5859375, + "learning_rate": 8.519860829225864e-06, + "loss": 0.134, + "step": 7920 + }, + { + "epoch": 1.1496140045667065, + "grad_norm": 1.28125, + "learning_rate": 8.505363873586548e-06, + "loss": 0.1348, + "step": 7930 + }, + { + "epoch": 1.151063752672973, + "grad_norm": 1.578125, + "learning_rate": 8.49086691794723e-06, + "loss": 0.1487, + "step": 7940 + }, + { + "epoch": 1.1525135007792395, + "grad_norm": 1.734375, + "learning_rate": 8.476369962307917e-06, + "loss": 0.1224, + "step": 7950 + }, + { + "epoch": 1.153963248885506, + "grad_norm": 1.1328125, + "learning_rate": 8.461873006668601e-06, + "loss": 0.1374, + "step": 7960 + }, + { + "epoch": 1.1554129969917728, + "grad_norm": 1.15625, + "learning_rate": 8.447376051029284e-06, + "loss": 0.1422, + "step": 7970 + }, + { + "epoch": 1.156862745098039, + "grad_norm": 1.3203125, + "learning_rate": 8.432879095389968e-06, + "loss": 0.1474, + "step": 7980 + }, + { + "epoch": 1.1583124932043058, + "grad_norm": 1.5390625, + "learning_rate": 8.418382139750653e-06, + "loss": 0.1357, + "step": 7990 + }, + { + "epoch": 1.1597622413105724, + "grad_norm": 1.9140625, + "learning_rate": 8.403885184111337e-06, + "loss": 0.1493, + "step": 8000 + }, + { + "epoch": 1.161211989416839, + "grad_norm": 1.625, + "learning_rate": 8.389388228472022e-06, + "loss": 0.1421, + "step": 8010 + }, + { + "epoch": 1.1626617375231054, + "grad_norm": 1.5078125, + "learning_rate": 8.374891272832706e-06, + "loss": 0.1398, + "step": 8020 + }, + { + "epoch": 1.164111485629372, + "grad_norm": 1.546875, + "learning_rate": 8.36039431719339e-06, + "loss": 0.1532, + "step": 8030 + }, + { + "epoch": 1.1655612337356385, + "grad_norm": 1.84375, + "learning_rate": 8.345897361554073e-06, + "loss": 0.1476, + "step": 8040 + }, + { + "epoch": 1.167010981841905, + "grad_norm": 1.8359375, + "learning_rate": 8.33140040591476e-06, + "loss": 0.1678, + "step": 8050 + }, + { + "epoch": 1.1684607299481715, + "grad_norm": 1.4140625, + "learning_rate": 8.316903450275444e-06, + "loss": 0.136, + "step": 8060 + }, + { + "epoch": 1.169910478054438, + "grad_norm": 1.8359375, + "learning_rate": 8.302406494636126e-06, + "loss": 0.1572, + "step": 8070 + }, + { + "epoch": 1.1713602261607046, + "grad_norm": 1.65625, + "learning_rate": 8.28790953899681e-06, + "loss": 0.1598, + "step": 8080 + }, + { + "epoch": 1.172809974266971, + "grad_norm": 1.734375, + "learning_rate": 8.273412583357495e-06, + "loss": 0.1707, + "step": 8090 + }, + { + "epoch": 1.1742597223732376, + "grad_norm": 1.5390625, + "learning_rate": 8.25891562771818e-06, + "loss": 0.15, + "step": 8100 + }, + { + "epoch": 1.1757094704795041, + "grad_norm": 1.7421875, + "learning_rate": 8.244418672078864e-06, + "loss": 0.1266, + "step": 8110 + }, + { + "epoch": 1.1771592185857707, + "grad_norm": 2.03125, + "learning_rate": 8.229921716439548e-06, + "loss": 0.1601, + "step": 8120 + }, + { + "epoch": 1.1786089666920372, + "grad_norm": 1.7109375, + "learning_rate": 8.215424760800233e-06, + "loss": 0.1898, + "step": 8130 + }, + { + "epoch": 1.1800587147983037, + "grad_norm": 1.8203125, + "learning_rate": 8.200927805160917e-06, + "loss": 0.1515, + "step": 8140 + }, + { + "epoch": 1.1815084629045702, + "grad_norm": 2.109375, + "learning_rate": 8.186430849521602e-06, + "loss": 0.1474, + "step": 8150 + }, + { + "epoch": 1.182958211010837, + "grad_norm": 1.6796875, + "learning_rate": 8.171933893882286e-06, + "loss": 0.138, + "step": 8160 + }, + { + "epoch": 1.1844079591171033, + "grad_norm": 1.609375, + "learning_rate": 8.157436938242969e-06, + "loss": 0.1262, + "step": 8170 + }, + { + "epoch": 1.18585770722337, + "grad_norm": 1.9609375, + "learning_rate": 8.142939982603653e-06, + "loss": 0.116, + "step": 8180 + }, + { + "epoch": 1.1873074553296366, + "grad_norm": 2.140625, + "learning_rate": 8.128443026964338e-06, + "loss": 0.1985, + "step": 8190 + }, + { + "epoch": 1.188757203435903, + "grad_norm": 1.5546875, + "learning_rate": 8.113946071325022e-06, + "loss": 0.1349, + "step": 8200 + }, + { + "epoch": 1.1902069515421696, + "grad_norm": 1.5859375, + "learning_rate": 8.099449115685707e-06, + "loss": 0.1422, + "step": 8210 + }, + { + "epoch": 1.1916566996484361, + "grad_norm": 1.7578125, + "learning_rate": 8.084952160046391e-06, + "loss": 0.1342, + "step": 8220 + }, + { + "epoch": 1.1931064477547026, + "grad_norm": 1.4765625, + "learning_rate": 8.070455204407075e-06, + "loss": 0.1449, + "step": 8230 + }, + { + "epoch": 1.1945561958609692, + "grad_norm": 1.53125, + "learning_rate": 8.05595824876776e-06, + "loss": 0.155, + "step": 8240 + }, + { + "epoch": 1.1960059439672357, + "grad_norm": 1.546875, + "learning_rate": 8.041461293128444e-06, + "loss": 0.1173, + "step": 8250 + }, + { + "epoch": 1.1974556920735022, + "grad_norm": 1.4140625, + "learning_rate": 8.026964337489129e-06, + "loss": 0.1455, + "step": 8260 + }, + { + "epoch": 1.1989054401797687, + "grad_norm": 1.65625, + "learning_rate": 8.012467381849813e-06, + "loss": 0.1417, + "step": 8270 + }, + { + "epoch": 1.2003551882860353, + "grad_norm": 2.234375, + "learning_rate": 7.997970426210496e-06, + "loss": 0.128, + "step": 8280 + }, + { + "epoch": 1.2018049363923018, + "grad_norm": 2.21875, + "learning_rate": 7.98347347057118e-06, + "loss": 0.1469, + "step": 8290 + }, + { + "epoch": 1.2032546844985683, + "grad_norm": 1.4140625, + "learning_rate": 7.968976514931866e-06, + "loss": 0.1509, + "step": 8300 + }, + { + "epoch": 1.2047044326048348, + "grad_norm": 1.6796875, + "learning_rate": 7.954479559292549e-06, + "loss": 0.1482, + "step": 8310 + }, + { + "epoch": 1.2061541807111014, + "grad_norm": 1.953125, + "learning_rate": 7.939982603653233e-06, + "loss": 0.1689, + "step": 8320 + }, + { + "epoch": 1.2076039288173679, + "grad_norm": 2.109375, + "learning_rate": 7.925485648013918e-06, + "loss": 0.1507, + "step": 8330 + }, + { + "epoch": 1.2090536769236344, + "grad_norm": 1.671875, + "learning_rate": 7.910988692374602e-06, + "loss": 0.1567, + "step": 8340 + }, + { + "epoch": 1.2105034250299012, + "grad_norm": 1.3125, + "learning_rate": 7.896491736735287e-06, + "loss": 0.1494, + "step": 8350 + }, + { + "epoch": 1.2119531731361677, + "grad_norm": 1.5390625, + "learning_rate": 7.881994781095971e-06, + "loss": 0.1451, + "step": 8360 + }, + { + "epoch": 1.2134029212424342, + "grad_norm": 1.6328125, + "learning_rate": 7.867497825456655e-06, + "loss": 0.1377, + "step": 8370 + }, + { + "epoch": 1.2148526693487007, + "grad_norm": 0.95703125, + "learning_rate": 7.853000869817338e-06, + "loss": 0.1404, + "step": 8380 + }, + { + "epoch": 1.2163024174549673, + "grad_norm": 1.859375, + "learning_rate": 7.838503914178023e-06, + "loss": 0.1345, + "step": 8390 + }, + { + "epoch": 1.2177521655612338, + "grad_norm": 1.4921875, + "learning_rate": 7.824006958538709e-06, + "loss": 0.1476, + "step": 8400 + }, + { + "epoch": 1.2192019136675003, + "grad_norm": 1.5078125, + "learning_rate": 7.809510002899391e-06, + "loss": 0.1515, + "step": 8410 + }, + { + "epoch": 1.2206516617737668, + "grad_norm": 1.484375, + "learning_rate": 7.795013047260076e-06, + "loss": 0.1585, + "step": 8420 + }, + { + "epoch": 1.2221014098800334, + "grad_norm": 1.3359375, + "learning_rate": 7.78051609162076e-06, + "loss": 0.1607, + "step": 8430 + }, + { + "epoch": 1.2235511579862999, + "grad_norm": 1.625, + "learning_rate": 7.766019135981445e-06, + "loss": 0.1474, + "step": 8440 + }, + { + "epoch": 1.2250009060925664, + "grad_norm": 1.625, + "learning_rate": 7.751522180342129e-06, + "loss": 0.1599, + "step": 8450 + }, + { + "epoch": 1.226450654198833, + "grad_norm": 2.40625, + "learning_rate": 7.737025224702814e-06, + "loss": 0.1479, + "step": 8460 + }, + { + "epoch": 1.2279004023050994, + "grad_norm": 1.328125, + "learning_rate": 7.722528269063498e-06, + "loss": 0.1627, + "step": 8470 + }, + { + "epoch": 1.229350150411366, + "grad_norm": 1.5859375, + "learning_rate": 7.70803131342418e-06, + "loss": 0.1603, + "step": 8480 + }, + { + "epoch": 1.2307998985176325, + "grad_norm": 1.8125, + "learning_rate": 7.693534357784865e-06, + "loss": 0.1657, + "step": 8490 + }, + { + "epoch": 1.232249646623899, + "grad_norm": 1.328125, + "learning_rate": 7.679037402145551e-06, + "loss": 0.119, + "step": 8500 + }, + { + "epoch": 1.2336993947301655, + "grad_norm": 2.203125, + "learning_rate": 7.664540446506234e-06, + "loss": 0.1333, + "step": 8510 + }, + { + "epoch": 1.2351491428364323, + "grad_norm": 1.59375, + "learning_rate": 7.650043490866918e-06, + "loss": 0.138, + "step": 8520 + }, + { + "epoch": 1.2365988909426986, + "grad_norm": 1.328125, + "learning_rate": 7.635546535227603e-06, + "loss": 0.1264, + "step": 8530 + }, + { + "epoch": 1.2380486390489653, + "grad_norm": 1.1484375, + "learning_rate": 7.621049579588286e-06, + "loss": 0.1392, + "step": 8540 + }, + { + "epoch": 1.2394983871552319, + "grad_norm": 1.7109375, + "learning_rate": 7.606552623948972e-06, + "loss": 0.1424, + "step": 8550 + }, + { + "epoch": 1.2409481352614984, + "grad_norm": 1.2890625, + "learning_rate": 7.592055668309656e-06, + "loss": 0.1755, + "step": 8560 + }, + { + "epoch": 1.242397883367765, + "grad_norm": 1.9296875, + "learning_rate": 7.57755871267034e-06, + "loss": 0.1673, + "step": 8570 + }, + { + "epoch": 1.2438476314740314, + "grad_norm": 1.734375, + "learning_rate": 7.563061757031024e-06, + "loss": 0.1711, + "step": 8580 + }, + { + "epoch": 1.245297379580298, + "grad_norm": 1.65625, + "learning_rate": 7.5485648013917076e-06, + "loss": 0.1388, + "step": 8590 + }, + { + "epoch": 1.2467471276865645, + "grad_norm": 1.484375, + "learning_rate": 7.534067845752393e-06, + "loss": 0.1503, + "step": 8600 + }, + { + "epoch": 1.248196875792831, + "grad_norm": 1.796875, + "learning_rate": 7.519570890113077e-06, + "loss": 0.1332, + "step": 8610 + }, + { + "epoch": 1.2496466238990975, + "grad_norm": 1.625, + "learning_rate": 7.505073934473761e-06, + "loss": 0.1447, + "step": 8620 + }, + { + "epoch": 1.251096372005364, + "grad_norm": 1.2578125, + "learning_rate": 7.490576978834445e-06, + "loss": 0.1505, + "step": 8630 + }, + { + "epoch": 1.2525461201116306, + "grad_norm": 2.0, + "learning_rate": 7.476080023195129e-06, + "loss": 0.156, + "step": 8640 + }, + { + "epoch": 1.253995868217897, + "grad_norm": 1.5234375, + "learning_rate": 7.461583067555814e-06, + "loss": 0.1516, + "step": 8650 + }, + { + "epoch": 1.2554456163241636, + "grad_norm": 1.8359375, + "learning_rate": 7.4470861119164985e-06, + "loss": 0.1396, + "step": 8660 + }, + { + "epoch": 1.2568953644304302, + "grad_norm": 1.6640625, + "learning_rate": 7.432589156277182e-06, + "loss": 0.122, + "step": 8670 + }, + { + "epoch": 1.2583451125366967, + "grad_norm": 1.3046875, + "learning_rate": 7.4180922006378665e-06, + "loss": 0.1394, + "step": 8680 + }, + { + "epoch": 1.2597948606429634, + "grad_norm": 1.4609375, + "learning_rate": 7.40359524499855e-06, + "loss": 0.133, + "step": 8690 + }, + { + "epoch": 1.2612446087492297, + "grad_norm": 1.6328125, + "learning_rate": 7.389098289359235e-06, + "loss": 0.148, + "step": 8700 + }, + { + "epoch": 1.2626943568554965, + "grad_norm": 1.8046875, + "learning_rate": 7.37460133371992e-06, + "loss": 0.1508, + "step": 8710 + }, + { + "epoch": 1.2641441049617628, + "grad_norm": 2.25, + "learning_rate": 7.360104378080603e-06, + "loss": 0.1478, + "step": 8720 + }, + { + "epoch": 1.2655938530680295, + "grad_norm": 1.6171875, + "learning_rate": 7.345607422441288e-06, + "loss": 0.1365, + "step": 8730 + }, + { + "epoch": 1.267043601174296, + "grad_norm": 1.203125, + "learning_rate": 7.331110466801971e-06, + "loss": 0.1482, + "step": 8740 + }, + { + "epoch": 1.2684933492805626, + "grad_norm": 1.5625, + "learning_rate": 7.3166135111626565e-06, + "loss": 0.1286, + "step": 8750 + }, + { + "epoch": 1.269943097386829, + "grad_norm": 1.3515625, + "learning_rate": 7.302116555523341e-06, + "loss": 0.146, + "step": 8760 + }, + { + "epoch": 1.2713928454930956, + "grad_norm": 1.65625, + "learning_rate": 7.2876195998840245e-06, + "loss": 0.1333, + "step": 8770 + }, + { + "epoch": 1.2728425935993621, + "grad_norm": 1.5390625, + "learning_rate": 7.273122644244709e-06, + "loss": 0.1275, + "step": 8780 + }, + { + "epoch": 1.2742923417056287, + "grad_norm": 1.796875, + "learning_rate": 7.258625688605393e-06, + "loss": 0.1324, + "step": 8790 + }, + { + "epoch": 1.2757420898118952, + "grad_norm": 1.2578125, + "learning_rate": 7.244128732966078e-06, + "loss": 0.1355, + "step": 8800 + }, + { + "epoch": 1.2771918379181617, + "grad_norm": 1.328125, + "learning_rate": 7.229631777326762e-06, + "loss": 0.1194, + "step": 8810 + }, + { + "epoch": 1.2786415860244282, + "grad_norm": 1.859375, + "learning_rate": 7.215134821687446e-06, + "loss": 0.1591, + "step": 8820 + }, + { + "epoch": 1.2800913341306948, + "grad_norm": 2.125, + "learning_rate": 7.20063786604813e-06, + "loss": 0.1346, + "step": 8830 + }, + { + "epoch": 1.2815410822369613, + "grad_norm": 1.59375, + "learning_rate": 7.186140910408815e-06, + "loss": 0.1511, + "step": 8840 + }, + { + "epoch": 1.2829908303432278, + "grad_norm": 2.4375, + "learning_rate": 7.171643954769499e-06, + "loss": 0.1494, + "step": 8850 + }, + { + "epoch": 1.2844405784494943, + "grad_norm": 1.3203125, + "learning_rate": 7.1571469991301834e-06, + "loss": 0.1316, + "step": 8860 + }, + { + "epoch": 1.2858903265557609, + "grad_norm": 1.7421875, + "learning_rate": 7.142650043490868e-06, + "loss": 0.1379, + "step": 8870 + }, + { + "epoch": 1.2873400746620276, + "grad_norm": 1.2421875, + "learning_rate": 7.128153087851551e-06, + "loss": 0.1206, + "step": 8880 + }, + { + "epoch": 1.288789822768294, + "grad_norm": 1.4375, + "learning_rate": 7.113656132212236e-06, + "loss": 0.1305, + "step": 8890 + }, + { + "epoch": 1.2902395708745606, + "grad_norm": 1.859375, + "learning_rate": 7.09915917657292e-06, + "loss": 0.1371, + "step": 8900 + }, + { + "epoch": 1.291689318980827, + "grad_norm": 1.2578125, + "learning_rate": 7.084662220933605e-06, + "loss": 0.1547, + "step": 8910 + }, + { + "epoch": 1.2931390670870937, + "grad_norm": 1.6640625, + "learning_rate": 7.070165265294289e-06, + "loss": 0.1494, + "step": 8920 + }, + { + "epoch": 1.2945888151933602, + "grad_norm": 1.53125, + "learning_rate": 7.055668309654973e-06, + "loss": 0.1254, + "step": 8930 + }, + { + "epoch": 1.2960385632996267, + "grad_norm": 1.8125, + "learning_rate": 7.041171354015657e-06, + "loss": 0.1467, + "step": 8940 + }, + { + "epoch": 1.2974883114058933, + "grad_norm": 2.1875, + "learning_rate": 7.026674398376342e-06, + "loss": 0.181, + "step": 8950 + }, + { + "epoch": 1.2989380595121598, + "grad_norm": 1.234375, + "learning_rate": 7.012177442737026e-06, + "loss": 0.1649, + "step": 8960 + }, + { + "epoch": 1.3003878076184263, + "grad_norm": 1.5703125, + "learning_rate": 6.99768048709771e-06, + "loss": 0.1485, + "step": 8970 + }, + { + "epoch": 1.3018375557246928, + "grad_norm": 1.578125, + "learning_rate": 6.983183531458394e-06, + "loss": 0.1288, + "step": 8980 + }, + { + "epoch": 1.3032873038309594, + "grad_norm": 1.84375, + "learning_rate": 6.968686575819078e-06, + "loss": 0.1448, + "step": 8990 + }, + { + "epoch": 1.304737051937226, + "grad_norm": 1.8359375, + "learning_rate": 6.9541896201797636e-06, + "loss": 0.1504, + "step": 9000 + }, + { + "epoch": 1.3061868000434924, + "grad_norm": 1.765625, + "learning_rate": 6.939692664540447e-06, + "loss": 0.1007, + "step": 9010 + }, + { + "epoch": 1.307636548149759, + "grad_norm": 1.4140625, + "learning_rate": 6.9251957089011316e-06, + "loss": 0.1324, + "step": 9020 + }, + { + "epoch": 1.3090862962560255, + "grad_norm": 1.078125, + "learning_rate": 6.910698753261815e-06, + "loss": 0.1452, + "step": 9030 + }, + { + "epoch": 1.310536044362292, + "grad_norm": 1.34375, + "learning_rate": 6.8962017976224995e-06, + "loss": 0.1224, + "step": 9040 + }, + { + "epoch": 1.3119857924685585, + "grad_norm": 1.2890625, + "learning_rate": 6.881704841983185e-06, + "loss": 0.1559, + "step": 9050 + }, + { + "epoch": 1.313435540574825, + "grad_norm": 1.7109375, + "learning_rate": 6.867207886343868e-06, + "loss": 0.151, + "step": 9060 + }, + { + "epoch": 1.3148852886810918, + "grad_norm": 1.765625, + "learning_rate": 6.852710930704553e-06, + "loss": 0.1558, + "step": 9070 + }, + { + "epoch": 1.316335036787358, + "grad_norm": 1.484375, + "learning_rate": 6.838213975065236e-06, + "loss": 0.137, + "step": 9080 + }, + { + "epoch": 1.3177847848936248, + "grad_norm": 1.90625, + "learning_rate": 6.823717019425921e-06, + "loss": 0.1493, + "step": 9090 + }, + { + "epoch": 1.3192345329998914, + "grad_norm": 1.3359375, + "learning_rate": 6.809220063786606e-06, + "loss": 0.136, + "step": 9100 + }, + { + "epoch": 1.3206842811061579, + "grad_norm": 1.6640625, + "learning_rate": 6.79472310814729e-06, + "loss": 0.156, + "step": 9110 + }, + { + "epoch": 1.3221340292124244, + "grad_norm": 1.265625, + "learning_rate": 6.780226152507974e-06, + "loss": 0.1402, + "step": 9120 + }, + { + "epoch": 1.323583777318691, + "grad_norm": 1.9375, + "learning_rate": 6.765729196868658e-06, + "loss": 0.1521, + "step": 9130 + }, + { + "epoch": 1.3250335254249574, + "grad_norm": 1.859375, + "learning_rate": 6.751232241229342e-06, + "loss": 0.1757, + "step": 9140 + }, + { + "epoch": 1.326483273531224, + "grad_norm": 1.5078125, + "learning_rate": 6.736735285590027e-06, + "loss": 0.13, + "step": 9150 + }, + { + "epoch": 1.3279330216374905, + "grad_norm": 1.4453125, + "learning_rate": 6.722238329950711e-06, + "loss": 0.1636, + "step": 9160 + }, + { + "epoch": 1.329382769743757, + "grad_norm": 1.65625, + "learning_rate": 6.707741374311395e-06, + "loss": 0.1518, + "step": 9170 + }, + { + "epoch": 1.3308325178500235, + "grad_norm": 1.5625, + "learning_rate": 6.693244418672079e-06, + "loss": 0.131, + "step": 9180 + }, + { + "epoch": 1.33228226595629, + "grad_norm": 1.6953125, + "learning_rate": 6.678747463032763e-06, + "loss": 0.1467, + "step": 9190 + }, + { + "epoch": 1.3337320140625566, + "grad_norm": 1.4765625, + "learning_rate": 6.6642505073934485e-06, + "loss": 0.142, + "step": 9200 + }, + { + "epoch": 1.3351817621688231, + "grad_norm": 2.125, + "learning_rate": 6.649753551754132e-06, + "loss": 0.1569, + "step": 9210 + }, + { + "epoch": 1.3366315102750896, + "grad_norm": 1.59375, + "learning_rate": 6.6352565961148165e-06, + "loss": 0.1396, + "step": 9220 + }, + { + "epoch": 1.3380812583813562, + "grad_norm": 1.828125, + "learning_rate": 6.6207596404755e-06, + "loss": 0.1422, + "step": 9230 + }, + { + "epoch": 1.339531006487623, + "grad_norm": 1.421875, + "learning_rate": 6.606262684836185e-06, + "loss": 0.1591, + "step": 9240 + }, + { + "epoch": 1.3409807545938892, + "grad_norm": 1.6796875, + "learning_rate": 6.59176572919687e-06, + "loss": 0.1591, + "step": 9250 + }, + { + "epoch": 1.342430502700156, + "grad_norm": 1.6015625, + "learning_rate": 6.577268773557553e-06, + "loss": 0.1325, + "step": 9260 + }, + { + "epoch": 1.3438802508064223, + "grad_norm": 1.6171875, + "learning_rate": 6.562771817918238e-06, + "loss": 0.1642, + "step": 9270 + }, + { + "epoch": 1.345329998912689, + "grad_norm": 1.765625, + "learning_rate": 6.548274862278921e-06, + "loss": 0.1454, + "step": 9280 + }, + { + "epoch": 1.3467797470189555, + "grad_norm": 1.8671875, + "learning_rate": 6.5337779066396066e-06, + "loss": 0.1538, + "step": 9290 + }, + { + "epoch": 1.348229495125222, + "grad_norm": 2.25, + "learning_rate": 6.519280951000291e-06, + "loss": 0.1572, + "step": 9300 + }, + { + "epoch": 1.3496792432314886, + "grad_norm": 1.703125, + "learning_rate": 6.5047839953609746e-06, + "loss": 0.1592, + "step": 9310 + }, + { + "epoch": 1.351128991337755, + "grad_norm": 1.421875, + "learning_rate": 6.490287039721659e-06, + "loss": 0.1377, + "step": 9320 + }, + { + "epoch": 1.3525787394440216, + "grad_norm": 1.421875, + "learning_rate": 6.4757900840823425e-06, + "loss": 0.1332, + "step": 9330 + }, + { + "epoch": 1.3540284875502882, + "grad_norm": 1.625, + "learning_rate": 6.461293128443028e-06, + "loss": 0.1433, + "step": 9340 + }, + { + "epoch": 1.3554782356565547, + "grad_norm": 1.1484375, + "learning_rate": 6.446796172803712e-06, + "loss": 0.1454, + "step": 9350 + }, + { + "epoch": 1.3569279837628212, + "grad_norm": 2.1875, + "learning_rate": 6.432299217164396e-06, + "loss": 0.1461, + "step": 9360 + }, + { + "epoch": 1.3583777318690877, + "grad_norm": 2.171875, + "learning_rate": 6.41780226152508e-06, + "loss": 0.1394, + "step": 9370 + }, + { + "epoch": 1.3598274799753542, + "grad_norm": 1.0234375, + "learning_rate": 6.403305305885764e-06, + "loss": 0.1265, + "step": 9380 + }, + { + "epoch": 1.3612772280816208, + "grad_norm": 1.625, + "learning_rate": 6.388808350246449e-06, + "loss": 0.1522, + "step": 9390 + }, + { + "epoch": 1.3627269761878873, + "grad_norm": 1.640625, + "learning_rate": 6.3743113946071335e-06, + "loss": 0.1219, + "step": 9400 + }, + { + "epoch": 1.3641767242941538, + "grad_norm": 1.375, + "learning_rate": 6.359814438967817e-06, + "loss": 0.128, + "step": 9410 + }, + { + "epoch": 1.3656264724004203, + "grad_norm": 1.6328125, + "learning_rate": 6.3453174833285014e-06, + "loss": 0.137, + "step": 9420 + }, + { + "epoch": 1.367076220506687, + "grad_norm": 1.359375, + "learning_rate": 6.330820527689185e-06, + "loss": 0.1358, + "step": 9430 + }, + { + "epoch": 1.3685259686129534, + "grad_norm": 2.28125, + "learning_rate": 6.31632357204987e-06, + "loss": 0.1379, + "step": 9440 + }, + { + "epoch": 1.3699757167192201, + "grad_norm": 2.265625, + "learning_rate": 6.301826616410555e-06, + "loss": 0.1302, + "step": 9450 + }, + { + "epoch": 1.3714254648254864, + "grad_norm": 1.8828125, + "learning_rate": 6.287329660771238e-06, + "loss": 0.1253, + "step": 9460 + }, + { + "epoch": 1.3728752129317532, + "grad_norm": 1.5234375, + "learning_rate": 6.272832705131923e-06, + "loss": 0.1174, + "step": 9470 + }, + { + "epoch": 1.3743249610380197, + "grad_norm": 1.765625, + "learning_rate": 6.258335749492607e-06, + "loss": 0.1542, + "step": 9480 + }, + { + "epoch": 1.3757747091442862, + "grad_norm": 2.03125, + "learning_rate": 6.2438387938532915e-06, + "loss": 0.1462, + "step": 9490 + }, + { + "epoch": 1.3772244572505528, + "grad_norm": 2.09375, + "learning_rate": 6.229341838213976e-06, + "loss": 0.135, + "step": 9500 + }, + { + "epoch": 1.3786742053568193, + "grad_norm": 1.5078125, + "learning_rate": 6.2148448825746595e-06, + "loss": 0.1356, + "step": 9510 + }, + { + "epoch": 1.3801239534630858, + "grad_norm": 1.875, + "learning_rate": 6.200347926935344e-06, + "loss": 0.1441, + "step": 9520 + }, + { + "epoch": 1.3815737015693523, + "grad_norm": 1.21875, + "learning_rate": 6.185850971296028e-06, + "loss": 0.1229, + "step": 9530 + }, + { + "epoch": 1.3830234496756189, + "grad_norm": 1.9921875, + "learning_rate": 6.171354015656713e-06, + "loss": 0.144, + "step": 9540 + }, + { + "epoch": 1.3844731977818854, + "grad_norm": 2.03125, + "learning_rate": 6.156857060017397e-06, + "loss": 0.17, + "step": 9550 + }, + { + "epoch": 1.385922945888152, + "grad_norm": 1.4765625, + "learning_rate": 6.142360104378082e-06, + "loss": 0.1334, + "step": 9560 + }, + { + "epoch": 1.3873726939944184, + "grad_norm": 1.8125, + "learning_rate": 6.127863148738765e-06, + "loss": 0.1346, + "step": 9570 + }, + { + "epoch": 1.388822442100685, + "grad_norm": 1.828125, + "learning_rate": 6.1133661930994496e-06, + "loss": 0.1415, + "step": 9580 + }, + { + "epoch": 1.3902721902069515, + "grad_norm": 1.9140625, + "learning_rate": 6.098869237460134e-06, + "loss": 0.1272, + "step": 9590 + }, + { + "epoch": 1.391721938313218, + "grad_norm": 1.625, + "learning_rate": 6.084372281820818e-06, + "loss": 0.15, + "step": 9600 + }, + { + "epoch": 1.3931716864194845, + "grad_norm": 1.59375, + "learning_rate": 6.069875326181503e-06, + "loss": 0.1199, + "step": 9610 + }, + { + "epoch": 1.3946214345257513, + "grad_norm": 1.4609375, + "learning_rate": 6.055378370542186e-06, + "loss": 0.1472, + "step": 9620 + }, + { + "epoch": 1.3960711826320176, + "grad_norm": 1.1328125, + "learning_rate": 6.040881414902871e-06, + "loss": 0.142, + "step": 9630 + }, + { + "epoch": 1.3975209307382843, + "grad_norm": 1.2421875, + "learning_rate": 6.026384459263555e-06, + "loss": 0.1429, + "step": 9640 + }, + { + "epoch": 1.3989706788445508, + "grad_norm": 1.984375, + "learning_rate": 6.01188750362424e-06, + "loss": 0.1489, + "step": 9650 + }, + { + "epoch": 1.4004204269508174, + "grad_norm": 1.6953125, + "learning_rate": 5.997390547984924e-06, + "loss": 0.1257, + "step": 9660 + }, + { + "epoch": 1.401870175057084, + "grad_norm": 1.7421875, + "learning_rate": 5.982893592345608e-06, + "loss": 0.1153, + "step": 9670 + }, + { + "epoch": 1.4033199231633504, + "grad_norm": 1.8515625, + "learning_rate": 5.968396636706292e-06, + "loss": 0.1309, + "step": 9680 + }, + { + "epoch": 1.404769671269617, + "grad_norm": 0.96875, + "learning_rate": 5.953899681066977e-06, + "loss": 0.1321, + "step": 9690 + }, + { + "epoch": 1.4062194193758835, + "grad_norm": 2.0625, + "learning_rate": 5.939402725427661e-06, + "loss": 0.1538, + "step": 9700 + }, + { + "epoch": 1.40766916748215, + "grad_norm": 1.4921875, + "learning_rate": 5.924905769788345e-06, + "loss": 0.1275, + "step": 9710 + }, + { + "epoch": 1.4091189155884165, + "grad_norm": 1.265625, + "learning_rate": 5.910408814149029e-06, + "loss": 0.1357, + "step": 9720 + }, + { + "epoch": 1.410568663694683, + "grad_norm": 1.703125, + "learning_rate": 5.895911858509713e-06, + "loss": 0.1476, + "step": 9730 + }, + { + "epoch": 1.4120184118009496, + "grad_norm": 1.75, + "learning_rate": 5.8814149028703985e-06, + "loss": 0.1362, + "step": 9740 + }, + { + "epoch": 1.413468159907216, + "grad_norm": 1.6484375, + "learning_rate": 5.866917947231082e-06, + "loss": 0.1374, + "step": 9750 + }, + { + "epoch": 1.4149179080134826, + "grad_norm": 2.1875, + "learning_rate": 5.8524209915917665e-06, + "loss": 0.1203, + "step": 9760 + }, + { + "epoch": 1.4163676561197491, + "grad_norm": 1.5234375, + "learning_rate": 5.83792403595245e-06, + "loss": 0.1219, + "step": 9770 + }, + { + "epoch": 1.4178174042260157, + "grad_norm": 1.8046875, + "learning_rate": 5.8234270803131345e-06, + "loss": 0.1472, + "step": 9780 + }, + { + "epoch": 1.4192671523322824, + "grad_norm": 1.1484375, + "learning_rate": 5.80893012467382e-06, + "loss": 0.1456, + "step": 9790 + }, + { + "epoch": 1.4207169004385487, + "grad_norm": 2.28125, + "learning_rate": 5.794433169034503e-06, + "loss": 0.1574, + "step": 9800 + }, + { + "epoch": 1.4221666485448154, + "grad_norm": 1.9296875, + "learning_rate": 5.779936213395188e-06, + "loss": 0.1487, + "step": 9810 + }, + { + "epoch": 1.4236163966510818, + "grad_norm": 1.3984375, + "learning_rate": 5.765439257755871e-06, + "loss": 0.1461, + "step": 9820 + }, + { + "epoch": 1.4250661447573485, + "grad_norm": 2.078125, + "learning_rate": 5.750942302116556e-06, + "loss": 0.1235, + "step": 9830 + }, + { + "epoch": 1.426515892863615, + "grad_norm": 2.1875, + "learning_rate": 5.736445346477241e-06, + "loss": 0.1512, + "step": 9840 + }, + { + "epoch": 1.4279656409698815, + "grad_norm": 1.5546875, + "learning_rate": 5.721948390837925e-06, + "loss": 0.1426, + "step": 9850 + }, + { + "epoch": 1.429415389076148, + "grad_norm": 2.5625, + "learning_rate": 5.707451435198609e-06, + "loss": 0.1434, + "step": 9860 + }, + { + "epoch": 1.4308651371824146, + "grad_norm": 2.046875, + "learning_rate": 5.6929544795592926e-06, + "loss": 0.1647, + "step": 9870 + }, + { + "epoch": 1.4323148852886811, + "grad_norm": 1.0078125, + "learning_rate": 5.678457523919977e-06, + "loss": 0.138, + "step": 9880 + }, + { + "epoch": 1.4337646333949476, + "grad_norm": 1.34375, + "learning_rate": 5.663960568280662e-06, + "loss": 0.1188, + "step": 9890 + }, + { + "epoch": 1.4352143815012142, + "grad_norm": 1.7109375, + "learning_rate": 5.649463612641346e-06, + "loss": 0.1315, + "step": 9900 + }, + { + "epoch": 1.4366641296074807, + "grad_norm": 1.8203125, + "learning_rate": 5.63496665700203e-06, + "loss": 0.1606, + "step": 9910 + }, + { + "epoch": 1.4381138777137472, + "grad_norm": 1.5234375, + "learning_rate": 5.620469701362714e-06, + "loss": 0.1469, + "step": 9920 + }, + { + "epoch": 1.4395636258200137, + "grad_norm": 1.4375, + "learning_rate": 5.605972745723398e-06, + "loss": 0.1186, + "step": 9930 + }, + { + "epoch": 1.4410133739262803, + "grad_norm": 1.9765625, + "learning_rate": 5.5914757900840835e-06, + "loss": 0.1299, + "step": 9940 + }, + { + "epoch": 1.4424631220325468, + "grad_norm": 1.484375, + "learning_rate": 5.576978834444767e-06, + "loss": 0.1372, + "step": 9950 + }, + { + "epoch": 1.4439128701388133, + "grad_norm": 1.296875, + "learning_rate": 5.5624818788054515e-06, + "loss": 0.1329, + "step": 9960 + }, + { + "epoch": 1.4453626182450798, + "grad_norm": 1.9375, + "learning_rate": 5.547984923166135e-06, + "loss": 0.1148, + "step": 9970 + }, + { + "epoch": 1.4468123663513466, + "grad_norm": 1.8671875, + "learning_rate": 5.5334879675268195e-06, + "loss": 0.1321, + "step": 9980 + }, + { + "epoch": 1.4482621144576129, + "grad_norm": 1.9296875, + "learning_rate": 5.518991011887505e-06, + "loss": 0.1128, + "step": 9990 + }, + { + "epoch": 1.4497118625638796, + "grad_norm": 1.8671875, + "learning_rate": 5.504494056248188e-06, + "loss": 0.1225, + "step": 10000 + }, + { + "epoch": 1.451161610670146, + "grad_norm": 1.5703125, + "learning_rate": 5.489997100608873e-06, + "loss": 0.1437, + "step": 10010 + }, + { + "epoch": 1.4526113587764127, + "grad_norm": 1.8984375, + "learning_rate": 5.475500144969556e-06, + "loss": 0.1649, + "step": 10020 + }, + { + "epoch": 1.4540611068826792, + "grad_norm": 1.71875, + "learning_rate": 5.461003189330241e-06, + "loss": 0.1737, + "step": 10030 + }, + { + "epoch": 1.4555108549889457, + "grad_norm": 1.90625, + "learning_rate": 5.446506233690926e-06, + "loss": 0.1581, + "step": 10040 + }, + { + "epoch": 1.4569606030952122, + "grad_norm": 1.515625, + "learning_rate": 5.4320092780516095e-06, + "loss": 0.1316, + "step": 10050 + }, + { + "epoch": 1.4584103512014788, + "grad_norm": 0.97265625, + "learning_rate": 5.417512322412294e-06, + "loss": 0.138, + "step": 10060 + }, + { + "epoch": 1.4598600993077453, + "grad_norm": 2.046875, + "learning_rate": 5.4030153667729775e-06, + "loss": 0.1498, + "step": 10070 + }, + { + "epoch": 1.4613098474140118, + "grad_norm": 1.3359375, + "learning_rate": 5.388518411133662e-06, + "loss": 0.1246, + "step": 10080 + }, + { + "epoch": 1.4627595955202783, + "grad_norm": 1.640625, + "learning_rate": 5.374021455494347e-06, + "loss": 0.1495, + "step": 10090 + }, + { + "epoch": 1.4642093436265449, + "grad_norm": 1.578125, + "learning_rate": 5.359524499855031e-06, + "loss": 0.127, + "step": 10100 + }, + { + "epoch": 1.4656590917328114, + "grad_norm": 2.28125, + "learning_rate": 5.345027544215715e-06, + "loss": 0.1651, + "step": 10110 + }, + { + "epoch": 1.467108839839078, + "grad_norm": 1.8203125, + "learning_rate": 5.330530588576399e-06, + "loss": 0.1323, + "step": 10120 + }, + { + "epoch": 1.4685585879453444, + "grad_norm": 2.1875, + "learning_rate": 5.316033632937083e-06, + "loss": 0.1609, + "step": 10130 + }, + { + "epoch": 1.470008336051611, + "grad_norm": 1.78125, + "learning_rate": 5.3015366772977684e-06, + "loss": 0.1559, + "step": 10140 + }, + { + "epoch": 1.4714580841578775, + "grad_norm": 2.1875, + "learning_rate": 5.287039721658452e-06, + "loss": 0.1457, + "step": 10150 + }, + { + "epoch": 1.472907832264144, + "grad_norm": 0.97265625, + "learning_rate": 5.2725427660191364e-06, + "loss": 0.1339, + "step": 10160 + }, + { + "epoch": 1.4743575803704108, + "grad_norm": 1.71875, + "learning_rate": 5.258045810379821e-06, + "loss": 0.1327, + "step": 10170 + }, + { + "epoch": 1.475807328476677, + "grad_norm": 1.4375, + "learning_rate": 5.243548854740504e-06, + "loss": 0.1456, + "step": 10180 + }, + { + "epoch": 1.4772570765829438, + "grad_norm": 1.3125, + "learning_rate": 5.22905189910119e-06, + "loss": 0.1408, + "step": 10190 + }, + { + "epoch": 1.4787068246892103, + "grad_norm": 1.4296875, + "learning_rate": 5.214554943461873e-06, + "loss": 0.1486, + "step": 10200 + }, + { + "epoch": 1.4801565727954769, + "grad_norm": 2.109375, + "learning_rate": 5.200057987822558e-06, + "loss": 0.1152, + "step": 10210 + }, + { + "epoch": 1.4816063209017434, + "grad_norm": 2.015625, + "learning_rate": 5.185561032183242e-06, + "loss": 0.1304, + "step": 10220 + }, + { + "epoch": 1.48305606900801, + "grad_norm": 1.578125, + "learning_rate": 5.171064076543926e-06, + "loss": 0.1556, + "step": 10230 + }, + { + "epoch": 1.4845058171142764, + "grad_norm": 1.84375, + "learning_rate": 5.156567120904611e-06, + "loss": 0.167, + "step": 10240 + }, + { + "epoch": 1.485955565220543, + "grad_norm": 1.6953125, + "learning_rate": 5.1420701652652945e-06, + "loss": 0.1518, + "step": 10250 + }, + { + "epoch": 1.4874053133268095, + "grad_norm": 1.5390625, + "learning_rate": 5.127573209625979e-06, + "loss": 0.1354, + "step": 10260 + }, + { + "epoch": 1.488855061433076, + "grad_norm": 1.5, + "learning_rate": 5.113076253986663e-06, + "loss": 0.1252, + "step": 10270 + }, + { + "epoch": 1.4903048095393425, + "grad_norm": 1.3828125, + "learning_rate": 5.098579298347347e-06, + "loss": 0.1391, + "step": 10280 + }, + { + "epoch": 1.491754557645609, + "grad_norm": 1.15625, + "learning_rate": 5.084082342708032e-06, + "loss": 0.1138, + "step": 10290 + }, + { + "epoch": 1.4932043057518756, + "grad_norm": 1.8828125, + "learning_rate": 5.0695853870687166e-06, + "loss": 0.1612, + "step": 10300 + }, + { + "epoch": 1.494654053858142, + "grad_norm": 2.078125, + "learning_rate": 5.0550884314294e-06, + "loss": 0.141, + "step": 10310 + }, + { + "epoch": 1.4961038019644086, + "grad_norm": 1.8515625, + "learning_rate": 5.0405914757900846e-06, + "loss": 0.1422, + "step": 10320 + }, + { + "epoch": 1.4975535500706751, + "grad_norm": 1.7578125, + "learning_rate": 5.026094520150768e-06, + "loss": 0.1298, + "step": 10330 + }, + { + "epoch": 1.499003298176942, + "grad_norm": 1.6796875, + "learning_rate": 5.011597564511453e-06, + "loss": 0.1318, + "step": 10340 + }, + { + "epoch": 1.5004530462832082, + "grad_norm": 1.796875, + "learning_rate": 4.997100608872138e-06, + "loss": 0.1643, + "step": 10350 + }, + { + "epoch": 1.501902794389475, + "grad_norm": 1.578125, + "learning_rate": 4.982603653232821e-06, + "loss": 0.1367, + "step": 10360 + }, + { + "epoch": 1.5033525424957412, + "grad_norm": 1.6015625, + "learning_rate": 4.968106697593506e-06, + "loss": 0.1388, + "step": 10370 + }, + { + "epoch": 1.504802290602008, + "grad_norm": 1.6171875, + "learning_rate": 4.95360974195419e-06, + "loss": 0.1498, + "step": 10380 + }, + { + "epoch": 1.5062520387082743, + "grad_norm": 1.875, + "learning_rate": 4.939112786314874e-06, + "loss": 0.1285, + "step": 10390 + }, + { + "epoch": 1.507701786814541, + "grad_norm": 1.28125, + "learning_rate": 4.924615830675559e-06, + "loss": 0.1401, + "step": 10400 + }, + { + "epoch": 1.5091515349208076, + "grad_norm": 1.65625, + "learning_rate": 4.910118875036243e-06, + "loss": 0.1459, + "step": 10410 + }, + { + "epoch": 1.510601283027074, + "grad_norm": 2.03125, + "learning_rate": 4.895621919396927e-06, + "loss": 0.1283, + "step": 10420 + }, + { + "epoch": 1.5120510311333406, + "grad_norm": 1.25, + "learning_rate": 4.8811249637576114e-06, + "loss": 0.1435, + "step": 10430 + }, + { + "epoch": 1.5135007792396071, + "grad_norm": 1.4921875, + "learning_rate": 4.866628008118295e-06, + "loss": 0.1174, + "step": 10440 + }, + { + "epoch": 1.5149505273458737, + "grad_norm": 1.53125, + "learning_rate": 4.85213105247898e-06, + "loss": 0.1323, + "step": 10450 + }, + { + "epoch": 1.5164002754521402, + "grad_norm": 1.6796875, + "learning_rate": 4.837634096839664e-06, + "loss": 0.1413, + "step": 10460 + }, + { + "epoch": 1.5178500235584067, + "grad_norm": 1.03125, + "learning_rate": 4.823137141200348e-06, + "loss": 0.1294, + "step": 10470 + }, + { + "epoch": 1.5192997716646732, + "grad_norm": 0.921875, + "learning_rate": 4.808640185561033e-06, + "loss": 0.1528, + "step": 10480 + }, + { + "epoch": 1.5207495197709398, + "grad_norm": 1.3671875, + "learning_rate": 4.794143229921716e-06, + "loss": 0.1331, + "step": 10490 + }, + { + "epoch": 1.5221992678772063, + "grad_norm": 1.4921875, + "learning_rate": 4.7796462742824015e-06, + "loss": 0.1311, + "step": 10500 + }, + { + "epoch": 1.523649015983473, + "grad_norm": 2.296875, + "learning_rate": 4.765149318643085e-06, + "loss": 0.1483, + "step": 10510 + }, + { + "epoch": 1.5250987640897393, + "grad_norm": 1.703125, + "learning_rate": 4.7506523630037695e-06, + "loss": 0.1527, + "step": 10520 + }, + { + "epoch": 1.526548512196006, + "grad_norm": 1.5703125, + "learning_rate": 4.736155407364454e-06, + "loss": 0.1228, + "step": 10530 + }, + { + "epoch": 1.5279982603022724, + "grad_norm": 1.890625, + "learning_rate": 4.721658451725138e-06, + "loss": 0.114, + "step": 10540 + }, + { + "epoch": 1.5294480084085391, + "grad_norm": 1.5078125, + "learning_rate": 4.707161496085823e-06, + "loss": 0.1416, + "step": 10550 + }, + { + "epoch": 1.5308977565148054, + "grad_norm": 1.5390625, + "learning_rate": 4.692664540446506e-06, + "loss": 0.1529, + "step": 10560 + }, + { + "epoch": 1.5323475046210722, + "grad_norm": 1.890625, + "learning_rate": 4.678167584807191e-06, + "loss": 0.1587, + "step": 10570 + }, + { + "epoch": 1.5337972527273385, + "grad_norm": 1.6328125, + "learning_rate": 4.663670629167875e-06, + "loss": 0.137, + "step": 10580 + }, + { + "epoch": 1.5352470008336052, + "grad_norm": 1.265625, + "learning_rate": 4.6491736735285596e-06, + "loss": 0.1449, + "step": 10590 + }, + { + "epoch": 1.5366967489398717, + "grad_norm": 1.8203125, + "learning_rate": 4.634676717889244e-06, + "loss": 0.122, + "step": 10600 + }, + { + "epoch": 1.5381464970461383, + "grad_norm": 1.8046875, + "learning_rate": 4.6201797622499276e-06, + "loss": 0.142, + "step": 10610 + }, + { + "epoch": 1.5395962451524048, + "grad_norm": 2.03125, + "learning_rate": 4.605682806610613e-06, + "loss": 0.1328, + "step": 10620 + }, + { + "epoch": 1.5410459932586713, + "grad_norm": 1.1796875, + "learning_rate": 4.591185850971296e-06, + "loss": 0.1382, + "step": 10630 + }, + { + "epoch": 1.5424957413649378, + "grad_norm": 1.7890625, + "learning_rate": 4.576688895331981e-06, + "loss": 0.1538, + "step": 10640 + }, + { + "epoch": 1.5439454894712044, + "grad_norm": 2.0625, + "learning_rate": 4.562191939692665e-06, + "loss": 0.1488, + "step": 10650 + }, + { + "epoch": 1.5453952375774709, + "grad_norm": 1.9765625, + "learning_rate": 4.547694984053349e-06, + "loss": 0.1379, + "step": 10660 + }, + { + "epoch": 1.5468449856837374, + "grad_norm": 1.8671875, + "learning_rate": 4.533198028414034e-06, + "loss": 0.1471, + "step": 10670 + }, + { + "epoch": 1.5482947337900042, + "grad_norm": 1.0078125, + "learning_rate": 4.518701072774718e-06, + "loss": 0.1446, + "step": 10680 + }, + { + "epoch": 1.5497444818962705, + "grad_norm": 1.6640625, + "learning_rate": 4.504204117135402e-06, + "loss": 0.1277, + "step": 10690 + }, + { + "epoch": 1.5511942300025372, + "grad_norm": 1.203125, + "learning_rate": 4.4897071614960865e-06, + "loss": 0.1534, + "step": 10700 + }, + { + "epoch": 1.5526439781088035, + "grad_norm": 1.515625, + "learning_rate": 4.47521020585677e-06, + "loss": 0.1328, + "step": 10710 + }, + { + "epoch": 1.5540937262150702, + "grad_norm": 2.234375, + "learning_rate": 4.460713250217455e-06, + "loss": 0.1544, + "step": 10720 + }, + { + "epoch": 1.5555434743213366, + "grad_norm": 1.671875, + "learning_rate": 4.446216294578139e-06, + "loss": 0.1258, + "step": 10730 + }, + { + "epoch": 1.5569932224276033, + "grad_norm": 1.890625, + "learning_rate": 4.431719338938823e-06, + "loss": 0.126, + "step": 10740 + }, + { + "epoch": 1.5584429705338696, + "grad_norm": 1.453125, + "learning_rate": 4.417222383299508e-06, + "loss": 0.1155, + "step": 10750 + }, + { + "epoch": 1.5598927186401363, + "grad_norm": 1.5, + "learning_rate": 4.402725427660191e-06, + "loss": 0.1544, + "step": 10760 + }, + { + "epoch": 1.5613424667464026, + "grad_norm": 2.609375, + "learning_rate": 4.3882284720208765e-06, + "loss": 0.1532, + "step": 10770 + }, + { + "epoch": 1.5627922148526694, + "grad_norm": 1.5, + "learning_rate": 4.37373151638156e-06, + "loss": 0.1228, + "step": 10780 + }, + { + "epoch": 1.564241962958936, + "grad_norm": 2.140625, + "learning_rate": 4.3592345607422445e-06, + "loss": 0.1487, + "step": 10790 + }, + { + "epoch": 1.5656917110652024, + "grad_norm": 1.4375, + "learning_rate": 4.344737605102929e-06, + "loss": 0.1615, + "step": 10800 + }, + { + "epoch": 1.567141459171469, + "grad_norm": 1.2890625, + "learning_rate": 4.3302406494636125e-06, + "loss": 0.1294, + "step": 10810 + }, + { + "epoch": 1.5685912072777355, + "grad_norm": 1.0703125, + "learning_rate": 4.315743693824298e-06, + "loss": 0.1238, + "step": 10820 + }, + { + "epoch": 1.570040955384002, + "grad_norm": 1.515625, + "learning_rate": 4.301246738184981e-06, + "loss": 0.1444, + "step": 10830 + }, + { + "epoch": 1.5714907034902685, + "grad_norm": 1.140625, + "learning_rate": 4.286749782545666e-06, + "loss": 0.1346, + "step": 10840 + }, + { + "epoch": 1.572940451596535, + "grad_norm": 1.6875, + "learning_rate": 4.27225282690635e-06, + "loss": 0.1442, + "step": 10850 + }, + { + "epoch": 1.5743901997028016, + "grad_norm": 1.6796875, + "learning_rate": 4.257755871267034e-06, + "loss": 0.1232, + "step": 10860 + }, + { + "epoch": 1.5758399478090683, + "grad_norm": 1.3125, + "learning_rate": 4.243258915627719e-06, + "loss": 0.132, + "step": 10870 + }, + { + "epoch": 1.5772896959153346, + "grad_norm": 1.5, + "learning_rate": 4.2287619599884026e-06, + "loss": 0.13, + "step": 10880 + }, + { + "epoch": 1.5787394440216014, + "grad_norm": 1.5078125, + "learning_rate": 4.214265004349087e-06, + "loss": 0.1422, + "step": 10890 + }, + { + "epoch": 1.5801891921278677, + "grad_norm": 2.09375, + "learning_rate": 4.199768048709771e-06, + "loss": 0.1499, + "step": 10900 + }, + { + "epoch": 1.5816389402341344, + "grad_norm": 2.078125, + "learning_rate": 4.185271093070456e-06, + "loss": 0.1508, + "step": 10910 + }, + { + "epoch": 1.5830886883404007, + "grad_norm": 1.875, + "learning_rate": 4.17077413743114e-06, + "loss": 0.1362, + "step": 10920 + }, + { + "epoch": 1.5845384364466675, + "grad_norm": 1.6640625, + "learning_rate": 4.156277181791824e-06, + "loss": 0.1366, + "step": 10930 + }, + { + "epoch": 1.5859881845529338, + "grad_norm": 1.984375, + "learning_rate": 4.141780226152508e-06, + "loss": 0.1423, + "step": 10940 + }, + { + "epoch": 1.5874379326592005, + "grad_norm": 1.8046875, + "learning_rate": 4.127283270513193e-06, + "loss": 0.1236, + "step": 10950 + }, + { + "epoch": 1.588887680765467, + "grad_norm": 1.5390625, + "learning_rate": 4.112786314873877e-06, + "loss": 0.1224, + "step": 10960 + }, + { + "epoch": 1.5903374288717336, + "grad_norm": 1.765625, + "learning_rate": 4.0982893592345615e-06, + "loss": 0.1482, + "step": 10970 + }, + { + "epoch": 1.591787176978, + "grad_norm": 1.8984375, + "learning_rate": 4.083792403595245e-06, + "loss": 0.1358, + "step": 10980 + }, + { + "epoch": 1.5932369250842666, + "grad_norm": 1.640625, + "learning_rate": 4.0692954479559295e-06, + "loss": 0.1401, + "step": 10990 + }, + { + "epoch": 1.5946866731905331, + "grad_norm": 1.828125, + "learning_rate": 4.054798492316614e-06, + "loss": 0.1464, + "step": 11000 + }, + { + "epoch": 1.5961364212967997, + "grad_norm": 1.828125, + "learning_rate": 4.040301536677298e-06, + "loss": 0.1673, + "step": 11010 + }, + { + "epoch": 1.5975861694030662, + "grad_norm": 1.1484375, + "learning_rate": 4.025804581037983e-06, + "loss": 0.1207, + "step": 11020 + }, + { + "epoch": 1.5990359175093327, + "grad_norm": 1.84375, + "learning_rate": 4.011307625398666e-06, + "loss": 0.1557, + "step": 11030 + }, + { + "epoch": 1.6004856656155992, + "grad_norm": 1.6328125, + "learning_rate": 3.996810669759351e-06, + "loss": 0.1062, + "step": 11040 + }, + { + "epoch": 1.6019354137218658, + "grad_norm": 1.4609375, + "learning_rate": 3.982313714120035e-06, + "loss": 0.1381, + "step": 11050 + }, + { + "epoch": 1.6033851618281325, + "grad_norm": 1.5234375, + "learning_rate": 3.9678167584807195e-06, + "loss": 0.1251, + "step": 11060 + }, + { + "epoch": 1.6048349099343988, + "grad_norm": 1.40625, + "learning_rate": 3.953319802841404e-06, + "loss": 0.1105, + "step": 11070 + }, + { + "epoch": 1.6062846580406656, + "grad_norm": 1.8125, + "learning_rate": 3.9388228472020875e-06, + "loss": 0.1526, + "step": 11080 + }, + { + "epoch": 1.6077344061469319, + "grad_norm": 1.625, + "learning_rate": 3.924325891562772e-06, + "loss": 0.1023, + "step": 11090 + }, + { + "epoch": 1.6091841542531986, + "grad_norm": 1.375, + "learning_rate": 3.909828935923456e-06, + "loss": 0.1256, + "step": 11100 + }, + { + "epoch": 1.610633902359465, + "grad_norm": 1.6171875, + "learning_rate": 3.895331980284141e-06, + "loss": 0.1316, + "step": 11110 + }, + { + "epoch": 1.6120836504657317, + "grad_norm": 1.5859375, + "learning_rate": 3.880835024644825e-06, + "loss": 0.1149, + "step": 11120 + }, + { + "epoch": 1.613533398571998, + "grad_norm": 1.9296875, + "learning_rate": 3.866338069005509e-06, + "loss": 0.1566, + "step": 11130 + }, + { + "epoch": 1.6149831466782647, + "grad_norm": 1.7421875, + "learning_rate": 3.851841113366193e-06, + "loss": 0.1409, + "step": 11140 + }, + { + "epoch": 1.6164328947845312, + "grad_norm": 1.984375, + "learning_rate": 3.837344157726878e-06, + "loss": 0.1386, + "step": 11150 + }, + { + "epoch": 1.6178826428907978, + "grad_norm": 1.78125, + "learning_rate": 3.822847202087562e-06, + "loss": 0.1384, + "step": 11160 + }, + { + "epoch": 1.6193323909970643, + "grad_norm": 1.0859375, + "learning_rate": 3.8083502464482464e-06, + "loss": 0.1484, + "step": 11170 + }, + { + "epoch": 1.6207821391033308, + "grad_norm": 1.6953125, + "learning_rate": 3.7938532908089304e-06, + "loss": 0.1436, + "step": 11180 + }, + { + "epoch": 1.6222318872095973, + "grad_norm": 1.703125, + "learning_rate": 3.7793563351696144e-06, + "loss": 0.1075, + "step": 11190 + }, + { + "epoch": 1.6236816353158638, + "grad_norm": 1.5625, + "learning_rate": 3.764859379530299e-06, + "loss": 0.1419, + "step": 11200 + }, + { + "epoch": 1.6251313834221304, + "grad_norm": 1.1875, + "learning_rate": 3.7503624238909832e-06, + "loss": 0.1185, + "step": 11210 + }, + { + "epoch": 1.626581131528397, + "grad_norm": 2.171875, + "learning_rate": 3.7358654682516677e-06, + "loss": 0.1326, + "step": 11220 + }, + { + "epoch": 1.6280308796346636, + "grad_norm": 1.609375, + "learning_rate": 3.7213685126123516e-06, + "loss": 0.1144, + "step": 11230 + }, + { + "epoch": 1.62948062774093, + "grad_norm": 2.234375, + "learning_rate": 3.7068715569730356e-06, + "loss": 0.161, + "step": 11240 + }, + { + "epoch": 1.6309303758471967, + "grad_norm": 1.828125, + "learning_rate": 3.6923746013337205e-06, + "loss": 0.1161, + "step": 11250 + }, + { + "epoch": 1.632380123953463, + "grad_norm": 1.7265625, + "learning_rate": 3.6778776456944045e-06, + "loss": 0.1494, + "step": 11260 + }, + { + "epoch": 1.6338298720597297, + "grad_norm": 1.359375, + "learning_rate": 3.663380690055089e-06, + "loss": 0.1405, + "step": 11270 + }, + { + "epoch": 1.635279620165996, + "grad_norm": 1.734375, + "learning_rate": 3.648883734415773e-06, + "loss": 0.12, + "step": 11280 + }, + { + "epoch": 1.6367293682722628, + "grad_norm": 1.6328125, + "learning_rate": 3.6343867787764577e-06, + "loss": 0.1451, + "step": 11290 + }, + { + "epoch": 1.638179116378529, + "grad_norm": 1.8671875, + "learning_rate": 3.6198898231371417e-06, + "loss": 0.1135, + "step": 11300 + }, + { + "epoch": 1.6396288644847958, + "grad_norm": 1.6953125, + "learning_rate": 3.6053928674978257e-06, + "loss": 0.1153, + "step": 11310 + }, + { + "epoch": 1.6410786125910624, + "grad_norm": 0.79296875, + "learning_rate": 3.59089591185851e-06, + "loss": 0.1381, + "step": 11320 + }, + { + "epoch": 1.6425283606973289, + "grad_norm": 1.5078125, + "learning_rate": 3.576398956219194e-06, + "loss": 0.1327, + "step": 11330 + }, + { + "epoch": 1.6439781088035954, + "grad_norm": 1.0546875, + "learning_rate": 3.561902000579879e-06, + "loss": 0.137, + "step": 11340 + }, + { + "epoch": 1.645427856909862, + "grad_norm": 1.8359375, + "learning_rate": 3.547405044940563e-06, + "loss": 0.1286, + "step": 11350 + }, + { + "epoch": 1.6468776050161285, + "grad_norm": 2.15625, + "learning_rate": 3.532908089301247e-06, + "loss": 0.1207, + "step": 11360 + }, + { + "epoch": 1.648327353122395, + "grad_norm": 1.7421875, + "learning_rate": 3.5184111336619314e-06, + "loss": 0.1441, + "step": 11370 + }, + { + "epoch": 1.6497771012286615, + "grad_norm": 1.6328125, + "learning_rate": 3.5039141780226154e-06, + "loss": 0.116, + "step": 11380 + }, + { + "epoch": 1.651226849334928, + "grad_norm": 2.15625, + "learning_rate": 3.4894172223833e-06, + "loss": 0.1413, + "step": 11390 + }, + { + "epoch": 1.6526765974411946, + "grad_norm": 1.6328125, + "learning_rate": 3.474920266743984e-06, + "loss": 0.1392, + "step": 11400 + }, + { + "epoch": 1.654126345547461, + "grad_norm": 1.5390625, + "learning_rate": 3.460423311104668e-06, + "loss": 0.1461, + "step": 11410 + }, + { + "epoch": 1.6555760936537278, + "grad_norm": 1.4921875, + "learning_rate": 3.4459263554653526e-06, + "loss": 0.1283, + "step": 11420 + }, + { + "epoch": 1.6570258417599941, + "grad_norm": 1.03125, + "learning_rate": 3.4314293998260366e-06, + "loss": 0.1343, + "step": 11430 + }, + { + "epoch": 1.6584755898662609, + "grad_norm": 1.640625, + "learning_rate": 3.4169324441867214e-06, + "loss": 0.1554, + "step": 11440 + }, + { + "epoch": 1.6599253379725272, + "grad_norm": 1.4296875, + "learning_rate": 3.4024354885474054e-06, + "loss": 0.1414, + "step": 11450 + }, + { + "epoch": 1.661375086078794, + "grad_norm": 1.6171875, + "learning_rate": 3.3879385329080894e-06, + "loss": 0.1368, + "step": 11460 + }, + { + "epoch": 1.6628248341850602, + "grad_norm": 2.625, + "learning_rate": 3.373441577268774e-06, + "loss": 0.1421, + "step": 11470 + }, + { + "epoch": 1.664274582291327, + "grad_norm": 1.953125, + "learning_rate": 3.358944621629458e-06, + "loss": 0.146, + "step": 11480 + }, + { + "epoch": 1.6657243303975933, + "grad_norm": 1.1484375, + "learning_rate": 3.3444476659901427e-06, + "loss": 0.1229, + "step": 11490 + }, + { + "epoch": 1.66717407850386, + "grad_norm": 1.3984375, + "learning_rate": 3.3299507103508267e-06, + "loss": 0.1294, + "step": 11500 + }, + { + "epoch": 1.6686238266101265, + "grad_norm": 1.2890625, + "learning_rate": 3.3154537547115107e-06, + "loss": 0.1153, + "step": 11510 + }, + { + "epoch": 1.670073574716393, + "grad_norm": 1.5703125, + "learning_rate": 3.300956799072195e-06, + "loss": 0.156, + "step": 11520 + }, + { + "epoch": 1.6715233228226596, + "grad_norm": 1.875, + "learning_rate": 3.286459843432879e-06, + "loss": 0.1423, + "step": 11530 + }, + { + "epoch": 1.672973070928926, + "grad_norm": 1.0234375, + "learning_rate": 3.271962887793564e-06, + "loss": 0.1286, + "step": 11540 + }, + { + "epoch": 1.6744228190351926, + "grad_norm": 1.375, + "learning_rate": 3.257465932154248e-06, + "loss": 0.1358, + "step": 11550 + }, + { + "epoch": 1.6758725671414592, + "grad_norm": 1.6484375, + "learning_rate": 3.242968976514932e-06, + "loss": 0.1352, + "step": 11560 + }, + { + "epoch": 1.6773223152477257, + "grad_norm": 1.421875, + "learning_rate": 3.2284720208756163e-06, + "loss": 0.154, + "step": 11570 + }, + { + "epoch": 1.6787720633539922, + "grad_norm": 1.7734375, + "learning_rate": 3.2139750652363007e-06, + "loss": 0.1541, + "step": 11580 + }, + { + "epoch": 1.6802218114602587, + "grad_norm": 1.265625, + "learning_rate": 3.199478109596985e-06, + "loss": 0.1346, + "step": 11590 + }, + { + "epoch": 1.6816715595665253, + "grad_norm": 1.8671875, + "learning_rate": 3.184981153957669e-06, + "loss": 0.1175, + "step": 11600 + }, + { + "epoch": 1.683121307672792, + "grad_norm": 1.8984375, + "learning_rate": 3.170484198318353e-06, + "loss": 0.1327, + "step": 11610 + }, + { + "epoch": 1.6845710557790583, + "grad_norm": 1.921875, + "learning_rate": 3.155987242679038e-06, + "loss": 0.1256, + "step": 11620 + }, + { + "epoch": 1.686020803885325, + "grad_norm": 2.0625, + "learning_rate": 3.141490287039722e-06, + "loss": 0.1291, + "step": 11630 + }, + { + "epoch": 1.6874705519915913, + "grad_norm": 2.15625, + "learning_rate": 3.1269933314004064e-06, + "loss": 0.1423, + "step": 11640 + }, + { + "epoch": 1.688920300097858, + "grad_norm": 1.8046875, + "learning_rate": 3.1124963757610904e-06, + "loss": 0.1358, + "step": 11650 + }, + { + "epoch": 1.6903700482041244, + "grad_norm": 0.87109375, + "learning_rate": 3.0979994201217744e-06, + "loss": 0.1407, + "step": 11660 + }, + { + "epoch": 1.6918197963103911, + "grad_norm": 2.34375, + "learning_rate": 3.083502464482459e-06, + "loss": 0.1258, + "step": 11670 + }, + { + "epoch": 1.6932695444166574, + "grad_norm": 0.98828125, + "learning_rate": 3.069005508843143e-06, + "loss": 0.1235, + "step": 11680 + }, + { + "epoch": 1.6947192925229242, + "grad_norm": 1.59375, + "learning_rate": 3.0545085532038276e-06, + "loss": 0.1391, + "step": 11690 + }, + { + "epoch": 1.6961690406291907, + "grad_norm": 1.0390625, + "learning_rate": 3.0400115975645116e-06, + "loss": 0.1025, + "step": 11700 + }, + { + "epoch": 1.6976187887354572, + "grad_norm": 1.59375, + "learning_rate": 3.0255146419251956e-06, + "loss": 0.1496, + "step": 11710 + }, + { + "epoch": 1.6990685368417238, + "grad_norm": 1.65625, + "learning_rate": 3.0110176862858804e-06, + "loss": 0.1331, + "step": 11720 + }, + { + "epoch": 1.7005182849479903, + "grad_norm": 1.96875, + "learning_rate": 2.9965207306465644e-06, + "loss": 0.1492, + "step": 11730 + }, + { + "epoch": 1.7019680330542568, + "grad_norm": 1.421875, + "learning_rate": 2.982023775007249e-06, + "loss": 0.1269, + "step": 11740 + }, + { + "epoch": 1.7034177811605233, + "grad_norm": 1.5234375, + "learning_rate": 2.967526819367933e-06, + "loss": 0.139, + "step": 11750 + }, + { + "epoch": 1.7048675292667899, + "grad_norm": 1.6328125, + "learning_rate": 2.953029863728617e-06, + "loss": 0.1191, + "step": 11760 + }, + { + "epoch": 1.7063172773730564, + "grad_norm": 1.84375, + "learning_rate": 2.9385329080893017e-06, + "loss": 0.1379, + "step": 11770 + }, + { + "epoch": 1.7077670254793231, + "grad_norm": 1.875, + "learning_rate": 2.9240359524499857e-06, + "loss": 0.1262, + "step": 11780 + }, + { + "epoch": 1.7092167735855894, + "grad_norm": 1.6953125, + "learning_rate": 2.90953899681067e-06, + "loss": 0.1282, + "step": 11790 + }, + { + "epoch": 1.7106665216918562, + "grad_norm": 1.671875, + "learning_rate": 2.895042041171354e-06, + "loss": 0.1232, + "step": 11800 + }, + { + "epoch": 1.7121162697981225, + "grad_norm": 1.6640625, + "learning_rate": 2.880545085532038e-06, + "loss": 0.1099, + "step": 11810 + }, + { + "epoch": 1.7135660179043892, + "grad_norm": 1.1796875, + "learning_rate": 2.866048129892723e-06, + "loss": 0.1261, + "step": 11820 + }, + { + "epoch": 1.7150157660106555, + "grad_norm": 1.2421875, + "learning_rate": 2.851551174253407e-06, + "loss": 0.1532, + "step": 11830 + }, + { + "epoch": 1.7164655141169223, + "grad_norm": 1.8515625, + "learning_rate": 2.8370542186140913e-06, + "loss": 0.1223, + "step": 11840 + }, + { + "epoch": 1.7179152622231886, + "grad_norm": 1.484375, + "learning_rate": 2.8225572629747753e-06, + "loss": 0.1288, + "step": 11850 + }, + { + "epoch": 1.7193650103294553, + "grad_norm": 2.265625, + "learning_rate": 2.8080603073354597e-06, + "loss": 0.1428, + "step": 11860 + }, + { + "epoch": 1.7208147584357218, + "grad_norm": 1.59375, + "learning_rate": 2.793563351696144e-06, + "loss": 0.1154, + "step": 11870 + }, + { + "epoch": 1.7222645065419884, + "grad_norm": 1.46875, + "learning_rate": 2.779066396056828e-06, + "loss": 0.1417, + "step": 11880 + }, + { + "epoch": 1.723714254648255, + "grad_norm": 1.859375, + "learning_rate": 2.7645694404175126e-06, + "loss": 0.1342, + "step": 11890 + }, + { + "epoch": 1.7251640027545214, + "grad_norm": 2.203125, + "learning_rate": 2.750072484778197e-06, + "loss": 0.1345, + "step": 11900 + }, + { + "epoch": 1.726613750860788, + "grad_norm": 1.65625, + "learning_rate": 2.735575529138881e-06, + "loss": 0.1484, + "step": 11910 + }, + { + "epoch": 1.7280634989670545, + "grad_norm": 2.125, + "learning_rate": 2.7210785734995654e-06, + "loss": 0.1257, + "step": 11920 + }, + { + "epoch": 1.729513247073321, + "grad_norm": 1.4375, + "learning_rate": 2.7065816178602494e-06, + "loss": 0.1381, + "step": 11930 + }, + { + "epoch": 1.7309629951795875, + "grad_norm": 1.5546875, + "learning_rate": 2.692084662220934e-06, + "loss": 0.1379, + "step": 11940 + }, + { + "epoch": 1.732412743285854, + "grad_norm": 1.546875, + "learning_rate": 2.6775877065816182e-06, + "loss": 0.1357, + "step": 11950 + }, + { + "epoch": 1.7338624913921206, + "grad_norm": 2.328125, + "learning_rate": 2.663090750942302e-06, + "loss": 0.1259, + "step": 11960 + }, + { + "epoch": 1.7353122394983873, + "grad_norm": 1.7265625, + "learning_rate": 2.6485937953029866e-06, + "loss": 0.1379, + "step": 11970 + }, + { + "epoch": 1.7367619876046536, + "grad_norm": 1.6953125, + "learning_rate": 2.6340968396636706e-06, + "loss": 0.1237, + "step": 11980 + }, + { + "epoch": 1.7382117357109204, + "grad_norm": 1.4921875, + "learning_rate": 2.6195998840243555e-06, + "loss": 0.1126, + "step": 11990 + }, + { + "epoch": 1.7396614838171867, + "grad_norm": 1.609375, + "learning_rate": 2.6051029283850395e-06, + "loss": 0.1129, + "step": 12000 + }, + { + "epoch": 1.7411112319234534, + "grad_norm": 1.6953125, + "learning_rate": 2.5906059727457234e-06, + "loss": 0.1497, + "step": 12010 + }, + { + "epoch": 1.7425609800297197, + "grad_norm": 1.453125, + "learning_rate": 2.576109017106408e-06, + "loss": 0.114, + "step": 12020 + }, + { + "epoch": 1.7440107281359865, + "grad_norm": 1.0, + "learning_rate": 2.561612061467092e-06, + "loss": 0.136, + "step": 12030 + }, + { + "epoch": 1.7454604762422528, + "grad_norm": 1.7265625, + "learning_rate": 2.5471151058277767e-06, + "loss": 0.1292, + "step": 12040 + }, + { + "epoch": 1.7469102243485195, + "grad_norm": 1.8359375, + "learning_rate": 2.5326181501884607e-06, + "loss": 0.1402, + "step": 12050 + }, + { + "epoch": 1.748359972454786, + "grad_norm": 1.5703125, + "learning_rate": 2.5181211945491447e-06, + "loss": 0.1286, + "step": 12060 + }, + { + "epoch": 1.7498097205610526, + "grad_norm": 2.3125, + "learning_rate": 2.503624238909829e-06, + "loss": 0.1481, + "step": 12070 + }, + { + "epoch": 1.751259468667319, + "grad_norm": 1.8671875, + "learning_rate": 2.4891272832705135e-06, + "loss": 0.1437, + "step": 12080 + }, + { + "epoch": 1.7527092167735856, + "grad_norm": 1.921875, + "learning_rate": 2.4746303276311975e-06, + "loss": 0.1494, + "step": 12090 + }, + { + "epoch": 1.7541589648798521, + "grad_norm": 1.6953125, + "learning_rate": 2.460133371991882e-06, + "loss": 0.1301, + "step": 12100 + }, + { + "epoch": 1.7556087129861186, + "grad_norm": 1.53125, + "learning_rate": 2.4456364163525663e-06, + "loss": 0.1361, + "step": 12110 + }, + { + "epoch": 1.7570584610923852, + "grad_norm": 1.6328125, + "learning_rate": 2.4311394607132503e-06, + "loss": 0.143, + "step": 12120 + }, + { + "epoch": 1.7585082091986517, + "grad_norm": 1.3203125, + "learning_rate": 2.4166425050739348e-06, + "loss": 0.1049, + "step": 12130 + }, + { + "epoch": 1.7599579573049182, + "grad_norm": 1.7265625, + "learning_rate": 2.4021455494346187e-06, + "loss": 0.1151, + "step": 12140 + }, + { + "epoch": 1.7614077054111847, + "grad_norm": 1.2578125, + "learning_rate": 2.387648593795303e-06, + "loss": 0.1508, + "step": 12150 + }, + { + "epoch": 1.7628574535174515, + "grad_norm": 1.28125, + "learning_rate": 2.3731516381559876e-06, + "loss": 0.1403, + "step": 12160 + }, + { + "epoch": 1.7643072016237178, + "grad_norm": 1.5625, + "learning_rate": 2.3586546825166716e-06, + "loss": 0.1281, + "step": 12170 + }, + { + "epoch": 1.7657569497299845, + "grad_norm": 1.296875, + "learning_rate": 2.344157726877356e-06, + "loss": 0.1475, + "step": 12180 + }, + { + "epoch": 1.7672066978362508, + "grad_norm": 1.625, + "learning_rate": 2.32966077123804e-06, + "loss": 0.1149, + "step": 12190 + }, + { + "epoch": 1.7686564459425176, + "grad_norm": 1.609375, + "learning_rate": 2.3151638155987244e-06, + "loss": 0.1483, + "step": 12200 + }, + { + "epoch": 1.7701061940487839, + "grad_norm": 1.46875, + "learning_rate": 2.300666859959409e-06, + "loss": 0.1201, + "step": 12210 + }, + { + "epoch": 1.7715559421550506, + "grad_norm": 1.390625, + "learning_rate": 2.286169904320093e-06, + "loss": 0.1396, + "step": 12220 + }, + { + "epoch": 1.773005690261317, + "grad_norm": 1.9609375, + "learning_rate": 2.2716729486807772e-06, + "loss": 0.1519, + "step": 12230 + }, + { + "epoch": 1.7744554383675837, + "grad_norm": 1.953125, + "learning_rate": 2.2571759930414612e-06, + "loss": 0.1423, + "step": 12240 + }, + { + "epoch": 1.7759051864738502, + "grad_norm": 1.7421875, + "learning_rate": 2.2426790374021456e-06, + "loss": 0.113, + "step": 12250 + }, + { + "epoch": 1.7773549345801167, + "grad_norm": 1.4140625, + "learning_rate": 2.22818208176283e-06, + "loss": 0.1269, + "step": 12260 + }, + { + "epoch": 1.7788046826863833, + "grad_norm": 0.96484375, + "learning_rate": 2.2136851261235145e-06, + "loss": 0.1281, + "step": 12270 + }, + { + "epoch": 1.7802544307926498, + "grad_norm": 2.015625, + "learning_rate": 2.1991881704841985e-06, + "loss": 0.1453, + "step": 12280 + }, + { + "epoch": 1.7817041788989163, + "grad_norm": 2.03125, + "learning_rate": 2.184691214844883e-06, + "loss": 0.1505, + "step": 12290 + }, + { + "epoch": 1.7831539270051828, + "grad_norm": 2.0, + "learning_rate": 2.170194259205567e-06, + "loss": 0.1416, + "step": 12300 + }, + { + "epoch": 1.7846036751114494, + "grad_norm": 1.7109375, + "learning_rate": 2.1556973035662513e-06, + "loss": 0.1319, + "step": 12310 + }, + { + "epoch": 1.7860534232177159, + "grad_norm": 1.78125, + "learning_rate": 2.1412003479269357e-06, + "loss": 0.1489, + "step": 12320 + }, + { + "epoch": 1.7875031713239826, + "grad_norm": 1.984375, + "learning_rate": 2.1267033922876197e-06, + "loss": 0.1276, + "step": 12330 + }, + { + "epoch": 1.788952919430249, + "grad_norm": 2.359375, + "learning_rate": 2.112206436648304e-06, + "loss": 0.1364, + "step": 12340 + }, + { + "epoch": 1.7904026675365157, + "grad_norm": 1.328125, + "learning_rate": 2.097709481008988e-06, + "loss": 0.1295, + "step": 12350 + }, + { + "epoch": 1.791852415642782, + "grad_norm": 1.5859375, + "learning_rate": 2.0832125253696725e-06, + "loss": 0.1025, + "step": 12360 + }, + { + "epoch": 1.7933021637490487, + "grad_norm": 2.015625, + "learning_rate": 2.068715569730357e-06, + "loss": 0.1587, + "step": 12370 + }, + { + "epoch": 1.794751911855315, + "grad_norm": 2.5625, + "learning_rate": 2.054218614091041e-06, + "loss": 0.1266, + "step": 12380 + }, + { + "epoch": 1.7962016599615818, + "grad_norm": 1.65625, + "learning_rate": 2.0397216584517254e-06, + "loss": 0.1226, + "step": 12390 + }, + { + "epoch": 1.797651408067848, + "grad_norm": 1.4296875, + "learning_rate": 2.0252247028124093e-06, + "loss": 0.1379, + "step": 12400 + }, + { + "epoch": 1.7991011561741148, + "grad_norm": 1.546875, + "learning_rate": 2.0107277471730938e-06, + "loss": 0.1476, + "step": 12410 + }, + { + "epoch": 1.8005509042803813, + "grad_norm": 1.328125, + "learning_rate": 1.996230791533778e-06, + "loss": 0.1109, + "step": 12420 + }, + { + "epoch": 1.8020006523866479, + "grad_norm": 1.8984375, + "learning_rate": 1.9817338358944626e-06, + "loss": 0.1235, + "step": 12430 + }, + { + "epoch": 1.8034504004929144, + "grad_norm": 2.015625, + "learning_rate": 1.9672368802551466e-06, + "loss": 0.1436, + "step": 12440 + }, + { + "epoch": 1.804900148599181, + "grad_norm": 1.8984375, + "learning_rate": 1.9527399246158306e-06, + "loss": 0.1263, + "step": 12450 + }, + { + "epoch": 1.8063498967054474, + "grad_norm": 1.7265625, + "learning_rate": 1.938242968976515e-06, + "loss": 0.136, + "step": 12460 + }, + { + "epoch": 1.807799644811714, + "grad_norm": 1.8515625, + "learning_rate": 1.9237460133371994e-06, + "loss": 0.1145, + "step": 12470 + }, + { + "epoch": 1.8092493929179805, + "grad_norm": 2.0, + "learning_rate": 1.909249057697884e-06, + "loss": 0.1444, + "step": 12480 + }, + { + "epoch": 1.810699141024247, + "grad_norm": 1.8125, + "learning_rate": 1.894752102058568e-06, + "loss": 0.1392, + "step": 12490 + }, + { + "epoch": 1.8121488891305135, + "grad_norm": 1.8515625, + "learning_rate": 1.880255146419252e-06, + "loss": 0.1706, + "step": 12500 + }, + { + "epoch": 1.81359863723678, + "grad_norm": 1.359375, + "learning_rate": 1.8657581907799362e-06, + "loss": 0.1197, + "step": 12510 + }, + { + "epoch": 1.8150483853430468, + "grad_norm": 2.03125, + "learning_rate": 1.8512612351406207e-06, + "loss": 0.1442, + "step": 12520 + }, + { + "epoch": 1.816498133449313, + "grad_norm": 1.84375, + "learning_rate": 1.8367642795013049e-06, + "loss": 0.1238, + "step": 12530 + }, + { + "epoch": 1.8179478815555798, + "grad_norm": 1.609375, + "learning_rate": 1.8222673238619893e-06, + "loss": 0.1096, + "step": 12540 + }, + { + "epoch": 1.8193976296618461, + "grad_norm": 2.125, + "learning_rate": 1.8077703682226733e-06, + "loss": 0.1416, + "step": 12550 + }, + { + "epoch": 1.820847377768113, + "grad_norm": 1.375, + "learning_rate": 1.7932734125833575e-06, + "loss": 0.1347, + "step": 12560 + }, + { + "epoch": 1.8222971258743792, + "grad_norm": 2.390625, + "learning_rate": 1.7787764569440419e-06, + "loss": 0.1701, + "step": 12570 + }, + { + "epoch": 1.823746873980646, + "grad_norm": 1.828125, + "learning_rate": 1.764279501304726e-06, + "loss": 0.1391, + "step": 12580 + }, + { + "epoch": 1.8251966220869122, + "grad_norm": 1.2734375, + "learning_rate": 1.7497825456654105e-06, + "loss": 0.1238, + "step": 12590 + }, + { + "epoch": 1.826646370193179, + "grad_norm": 2.1875, + "learning_rate": 1.7352855900260945e-06, + "loss": 0.143, + "step": 12600 + }, + { + "epoch": 1.8280961182994455, + "grad_norm": 1.59375, + "learning_rate": 1.720788634386779e-06, + "loss": 0.1337, + "step": 12610 + }, + { + "epoch": 1.829545866405712, + "grad_norm": 1.84375, + "learning_rate": 1.7062916787474631e-06, + "loss": 0.1133, + "step": 12620 + }, + { + "epoch": 1.8309956145119786, + "grad_norm": 1.4609375, + "learning_rate": 1.6917947231081475e-06, + "loss": 0.1237, + "step": 12630 + }, + { + "epoch": 1.832445362618245, + "grad_norm": 2.09375, + "learning_rate": 1.6772977674688317e-06, + "loss": 0.1576, + "step": 12640 + }, + { + "epoch": 1.8338951107245116, + "grad_norm": 2.5625, + "learning_rate": 1.6628008118295157e-06, + "loss": 0.1394, + "step": 12650 + }, + { + "epoch": 1.8353448588307781, + "grad_norm": 1.65625, + "learning_rate": 1.6483038561902002e-06, + "loss": 0.1366, + "step": 12660 + }, + { + "epoch": 1.8367946069370447, + "grad_norm": 2.171875, + "learning_rate": 1.6338069005508844e-06, + "loss": 0.1576, + "step": 12670 + }, + { + "epoch": 1.8382443550433112, + "grad_norm": 2.296875, + "learning_rate": 1.6193099449115688e-06, + "loss": 0.1356, + "step": 12680 + }, + { + "epoch": 1.8396941031495777, + "grad_norm": 1.28125, + "learning_rate": 1.604812989272253e-06, + "loss": 0.135, + "step": 12690 + }, + { + "epoch": 1.8411438512558442, + "grad_norm": 1.390625, + "learning_rate": 1.5903160336329374e-06, + "loss": 0.1101, + "step": 12700 + }, + { + "epoch": 1.842593599362111, + "grad_norm": 1.5078125, + "learning_rate": 1.5758190779936214e-06, + "loss": 0.1127, + "step": 12710 + }, + { + "epoch": 1.8440433474683773, + "grad_norm": 1.3671875, + "learning_rate": 1.5613221223543056e-06, + "loss": 0.1302, + "step": 12720 + }, + { + "epoch": 1.845493095574644, + "grad_norm": 1.8984375, + "learning_rate": 1.54682516671499e-06, + "loss": 0.1287, + "step": 12730 + }, + { + "epoch": 1.8469428436809103, + "grad_norm": 1.5859375, + "learning_rate": 1.5323282110756742e-06, + "loss": 0.1195, + "step": 12740 + }, + { + "epoch": 1.848392591787177, + "grad_norm": 2.28125, + "learning_rate": 1.5178312554363586e-06, + "loss": 0.1354, + "step": 12750 + }, + { + "epoch": 1.8498423398934434, + "grad_norm": 1.9765625, + "learning_rate": 1.5033342997970426e-06, + "loss": 0.1462, + "step": 12760 + }, + { + "epoch": 1.8512920879997101, + "grad_norm": 2.6875, + "learning_rate": 1.488837344157727e-06, + "loss": 0.1494, + "step": 12770 + }, + { + "epoch": 1.8527418361059764, + "grad_norm": 1.1640625, + "learning_rate": 1.4743403885184113e-06, + "loss": 0.1612, + "step": 12780 + }, + { + "epoch": 1.8541915842122432, + "grad_norm": 1.46875, + "learning_rate": 1.4598434328790957e-06, + "loss": 0.1368, + "step": 12790 + }, + { + "epoch": 1.8556413323185097, + "grad_norm": 1.890625, + "learning_rate": 1.4453464772397799e-06, + "loss": 0.1053, + "step": 12800 + }, + { + "epoch": 1.8570910804247762, + "grad_norm": 1.9921875, + "learning_rate": 1.4308495216004639e-06, + "loss": 0.142, + "step": 12810 + }, + { + "epoch": 1.8585408285310427, + "grad_norm": 1.6171875, + "learning_rate": 1.4163525659611483e-06, + "loss": 0.1275, + "step": 12820 + }, + { + "epoch": 1.8599905766373093, + "grad_norm": 1.265625, + "learning_rate": 1.4018556103218325e-06, + "loss": 0.1309, + "step": 12830 + }, + { + "epoch": 1.8614403247435758, + "grad_norm": 0.7890625, + "learning_rate": 1.387358654682517e-06, + "loss": 0.1232, + "step": 12840 + }, + { + "epoch": 1.8628900728498423, + "grad_norm": 1.515625, + "learning_rate": 1.3728616990432011e-06, + "loss": 0.0871, + "step": 12850 + }, + { + "epoch": 1.8643398209561088, + "grad_norm": 1.796875, + "learning_rate": 1.358364743403885e-06, + "loss": 0.1547, + "step": 12860 + }, + { + "epoch": 1.8657895690623754, + "grad_norm": 1.7265625, + "learning_rate": 1.3438677877645695e-06, + "loss": 0.1127, + "step": 12870 + }, + { + "epoch": 1.867239317168642, + "grad_norm": 1.6640625, + "learning_rate": 1.3293708321252537e-06, + "loss": 0.1325, + "step": 12880 + }, + { + "epoch": 1.8686890652749084, + "grad_norm": 1.640625, + "learning_rate": 1.3148738764859381e-06, + "loss": 0.1103, + "step": 12890 + }, + { + "epoch": 1.8701388133811752, + "grad_norm": 1.859375, + "learning_rate": 1.3003769208466223e-06, + "loss": 0.1392, + "step": 12900 + }, + { + "epoch": 1.8715885614874415, + "grad_norm": 1.0546875, + "learning_rate": 1.2858799652073066e-06, + "loss": 0.1289, + "step": 12910 + }, + { + "epoch": 1.8730383095937082, + "grad_norm": 0.98828125, + "learning_rate": 1.2713830095679908e-06, + "loss": 0.1137, + "step": 12920 + }, + { + "epoch": 1.8744880576999745, + "grad_norm": 1.3984375, + "learning_rate": 1.2568860539286752e-06, + "loss": 0.1323, + "step": 12930 + }, + { + "epoch": 1.8759378058062413, + "grad_norm": 1.5859375, + "learning_rate": 1.2423890982893594e-06, + "loss": 0.1294, + "step": 12940 + }, + { + "epoch": 1.8773875539125076, + "grad_norm": 0.98828125, + "learning_rate": 1.2278921426500436e-06, + "loss": 0.1154, + "step": 12950 + }, + { + "epoch": 1.8788373020187743, + "grad_norm": 1.0625, + "learning_rate": 1.2133951870107278e-06, + "loss": 0.122, + "step": 12960 + }, + { + "epoch": 1.8802870501250408, + "grad_norm": 2.328125, + "learning_rate": 1.198898231371412e-06, + "loss": 0.1568, + "step": 12970 + }, + { + "epoch": 1.8817367982313074, + "grad_norm": 1.75, + "learning_rate": 1.1844012757320964e-06, + "loss": 0.1479, + "step": 12980 + }, + { + "epoch": 1.8831865463375739, + "grad_norm": 2.109375, + "learning_rate": 1.1699043200927806e-06, + "loss": 0.1343, + "step": 12990 + }, + { + "epoch": 1.8846362944438404, + "grad_norm": 1.5, + "learning_rate": 1.1554073644534648e-06, + "loss": 0.1237, + "step": 13000 + }, + { + "epoch": 1.886086042550107, + "grad_norm": 1.7265625, + "learning_rate": 1.1409104088141492e-06, + "loss": 0.1253, + "step": 13010 + }, + { + "epoch": 1.8875357906563734, + "grad_norm": 1.4375, + "learning_rate": 1.1264134531748334e-06, + "loss": 0.1332, + "step": 13020 + }, + { + "epoch": 1.88898553876264, + "grad_norm": 1.3203125, + "learning_rate": 1.1119164975355176e-06, + "loss": 0.1179, + "step": 13030 + }, + { + "epoch": 1.8904352868689065, + "grad_norm": 1.1328125, + "learning_rate": 1.0974195418962019e-06, + "loss": 0.134, + "step": 13040 + }, + { + "epoch": 1.891885034975173, + "grad_norm": 2.15625, + "learning_rate": 1.082922586256886e-06, + "loss": 0.1253, + "step": 13050 + }, + { + "epoch": 1.8933347830814395, + "grad_norm": 1.921875, + "learning_rate": 1.0684256306175705e-06, + "loss": 0.1246, + "step": 13060 + }, + { + "epoch": 1.8947845311877063, + "grad_norm": 1.8671875, + "learning_rate": 1.0539286749782547e-06, + "loss": 0.1266, + "step": 13070 + }, + { + "epoch": 1.8962342792939726, + "grad_norm": 1.421875, + "learning_rate": 1.0394317193389389e-06, + "loss": 0.1374, + "step": 13080 + }, + { + "epoch": 1.8976840274002393, + "grad_norm": 1.625, + "learning_rate": 1.024934763699623e-06, + "loss": 0.1556, + "step": 13090 + }, + { + "epoch": 1.8991337755065056, + "grad_norm": 1.8828125, + "learning_rate": 1.0104378080603073e-06, + "loss": 0.1232, + "step": 13100 + }, + { + "epoch": 1.9005835236127724, + "grad_norm": 1.390625, + "learning_rate": 9.959408524209917e-07, + "loss": 0.1189, + "step": 13110 + }, + { + "epoch": 1.9020332717190387, + "grad_norm": 2.109375, + "learning_rate": 9.81443896781676e-07, + "loss": 0.1498, + "step": 13120 + }, + { + "epoch": 1.9034830198253054, + "grad_norm": 1.2734375, + "learning_rate": 9.669469411423601e-07, + "loss": 0.1316, + "step": 13130 + }, + { + "epoch": 1.9049327679315717, + "grad_norm": 2.40625, + "learning_rate": 9.524499855030444e-07, + "loss": 0.1414, + "step": 13140 + }, + { + "epoch": 1.9063825160378385, + "grad_norm": 2.046875, + "learning_rate": 9.379530298637286e-07, + "loss": 0.1181, + "step": 13150 + }, + { + "epoch": 1.907832264144105, + "grad_norm": 1.25, + "learning_rate": 9.234560742244129e-07, + "loss": 0.1222, + "step": 13160 + }, + { + "epoch": 1.9092820122503715, + "grad_norm": 1.6875, + "learning_rate": 9.089591185850973e-07, + "loss": 0.1212, + "step": 13170 + }, + { + "epoch": 1.910731760356638, + "grad_norm": 1.234375, + "learning_rate": 8.944621629457815e-07, + "loss": 0.1413, + "step": 13180 + }, + { + "epoch": 1.9121815084629046, + "grad_norm": 1.8828125, + "learning_rate": 8.799652073064658e-07, + "loss": 0.1391, + "step": 13190 + }, + { + "epoch": 1.913631256569171, + "grad_norm": 1.4921875, + "learning_rate": 8.654682516671499e-07, + "loss": 0.1182, + "step": 13200 + }, + { + "epoch": 1.9150810046754376, + "grad_norm": 1.375, + "learning_rate": 8.509712960278342e-07, + "loss": 0.1083, + "step": 13210 + }, + { + "epoch": 1.9165307527817041, + "grad_norm": 1.765625, + "learning_rate": 8.364743403885185e-07, + "loss": 0.135, + "step": 13220 + }, + { + "epoch": 1.9179805008879707, + "grad_norm": 2.0, + "learning_rate": 8.219773847492027e-07, + "loss": 0.1493, + "step": 13230 + }, + { + "epoch": 1.9194302489942372, + "grad_norm": 2.0, + "learning_rate": 8.07480429109887e-07, + "loss": 0.1372, + "step": 13240 + }, + { + "epoch": 1.9208799971005037, + "grad_norm": 2.046875, + "learning_rate": 7.929834734705713e-07, + "loss": 0.1408, + "step": 13250 + }, + { + "epoch": 1.9223297452067705, + "grad_norm": 1.2421875, + "learning_rate": 7.784865178312555e-07, + "loss": 0.1256, + "step": 13260 + }, + { + "epoch": 1.9237794933130368, + "grad_norm": 1.8125, + "learning_rate": 7.639895621919398e-07, + "loss": 0.1378, + "step": 13270 + }, + { + "epoch": 1.9252292414193035, + "grad_norm": 1.90625, + "learning_rate": 7.494926065526239e-07, + "loss": 0.1188, + "step": 13280 + }, + { + "epoch": 1.9266789895255698, + "grad_norm": 2.03125, + "learning_rate": 7.349956509133082e-07, + "loss": 0.1508, + "step": 13290 + }, + { + "epoch": 1.9281287376318366, + "grad_norm": 1.7734375, + "learning_rate": 7.204986952739926e-07, + "loss": 0.1231, + "step": 13300 + }, + { + "epoch": 1.9295784857381029, + "grad_norm": 1.9375, + "learning_rate": 7.060017396346768e-07, + "loss": 0.1388, + "step": 13310 + }, + { + "epoch": 1.9310282338443696, + "grad_norm": 1.53125, + "learning_rate": 6.915047839953611e-07, + "loss": 0.1279, + "step": 13320 + }, + { + "epoch": 1.932477981950636, + "grad_norm": 1.375, + "learning_rate": 6.770078283560453e-07, + "loss": 0.139, + "step": 13330 + }, + { + "epoch": 1.9339277300569027, + "grad_norm": 1.828125, + "learning_rate": 6.625108727167296e-07, + "loss": 0.1505, + "step": 13340 + }, + { + "epoch": 1.9353774781631692, + "grad_norm": 1.78125, + "learning_rate": 6.480139170774138e-07, + "loss": 0.1573, + "step": 13350 + }, + { + "epoch": 1.9368272262694357, + "grad_norm": 1.7578125, + "learning_rate": 6.33516961438098e-07, + "loss": 0.1416, + "step": 13360 + }, + { + "epoch": 1.9382769743757022, + "grad_norm": 1.8515625, + "learning_rate": 6.190200057987823e-07, + "loss": 0.1391, + "step": 13370 + }, + { + "epoch": 1.9397267224819688, + "grad_norm": 1.515625, + "learning_rate": 6.045230501594665e-07, + "loss": 0.1382, + "step": 13380 + }, + { + "epoch": 1.9411764705882353, + "grad_norm": 2.09375, + "learning_rate": 5.900260945201508e-07, + "loss": 0.1424, + "step": 13390 + }, + { + "epoch": 1.9426262186945018, + "grad_norm": 1.28125, + "learning_rate": 5.75529138880835e-07, + "loss": 0.1134, + "step": 13400 + }, + { + "epoch": 1.9440759668007683, + "grad_norm": 0.9140625, + "learning_rate": 5.610321832415193e-07, + "loss": 0.124, + "step": 13410 + }, + { + "epoch": 1.9455257149070349, + "grad_norm": 2.0625, + "learning_rate": 5.465352276022035e-07, + "loss": 0.1528, + "step": 13420 + }, + { + "epoch": 1.9469754630133016, + "grad_norm": 1.53125, + "learning_rate": 5.320382719628879e-07, + "loss": 0.1109, + "step": 13430 + }, + { + "epoch": 1.948425211119568, + "grad_norm": 1.8203125, + "learning_rate": 5.175413163235721e-07, + "loss": 0.1667, + "step": 13440 + }, + { + "epoch": 1.9498749592258346, + "grad_norm": 1.984375, + "learning_rate": 5.030443606842563e-07, + "loss": 0.1352, + "step": 13450 + }, + { + "epoch": 1.951324707332101, + "grad_norm": 1.7578125, + "learning_rate": 4.885474050449406e-07, + "loss": 0.1284, + "step": 13460 + }, + { + "epoch": 1.9527744554383677, + "grad_norm": 1.2734375, + "learning_rate": 4.740504494056249e-07, + "loss": 0.1143, + "step": 13470 + }, + { + "epoch": 1.954224203544634, + "grad_norm": 1.90625, + "learning_rate": 4.595534937663091e-07, + "loss": 0.1445, + "step": 13480 + }, + { + "epoch": 1.9556739516509007, + "grad_norm": 1.4296875, + "learning_rate": 4.4505653812699335e-07, + "loss": 0.1381, + "step": 13490 + }, + { + "epoch": 1.957123699757167, + "grad_norm": 1.4921875, + "learning_rate": 4.3055958248767766e-07, + "loss": 0.1422, + "step": 13500 + } + ], + "logging_steps": 10, + "max_steps": 13796, + "num_input_tokens_seen": 0, + "num_train_epochs": 2, + "save_steps": 500, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 5.825344102534742e+18, + "train_batch_size": 1, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-13500/training_args.bin b/checkpoint-13500/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..06838b874d6df8d2cf7fbec1647b56bbeee05efb --- /dev/null +++ b/checkpoint-13500/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0b692ac805c6882ecc47c1ef2d34907d8e095f209448815d8db422f0c5955b6d +size 5777 diff --git a/checkpoint-13796/chat_template.jinja b/checkpoint-13796/chat_template.jinja new file mode 100644 index 0000000000000000000000000000000000000000..ed10d0cf20ed185c105baf9a5234ab076f0b4803 --- /dev/null +++ b/checkpoint-13796/chat_template.jinja @@ -0,0 +1,37 @@ +[gMASK]{% for item in messages %}{% if item['tools'] is defined %}<|system|> +你是一个名为 ChatGLM 的人工智能助手。你是基于智谱AI训练的语言模型 GLM-4 模型开发的，你的任务是针对用户的问题和要求提供适当的答复和支持。 + +# 可用工具{% set tools = item['tools'] %}{% for tool in tools %}{% if tool['type'] == 'function' %} + +## {{ tool['function']['name'] }} + +{{ tool['function'] | tojson(indent=4) }} +在调用上述函数时，请使用 Json 格式表示调用的参数。{% elif tool['type'] == 'python' %} + +## python + +当你向 `python` 发送包含 Python 代码的消息时，该代码将会在一个有状态的 Jupyter notebook 环境中执行。 +`python` 返回代码执行的输出，或在执行 60 秒后返回超时。 +`/mnt/data` 将会持久化存储你的文件。在此会话中，`python` 无法访问互联网。不要使用 `python` 进行任何网络请求或者在线 API 调用，这些在线内容的访问将不会成功。{% elif tool['type'] == 'simple_browser' %} + +## simple_browser + +你可以使用 `simple_browser` 工具。该工具支持以下函数： +`search(query: str, recency_days: int)`：使用搜索引擎进行查询并显示结果，可以使用 `recency_days` 参数控制搜索内容的时效性。 +`mclick(ids: list[int])`：获取一系列指定 id 的页面内容。每次调用时，须选择3-10个页面。选择多个角度的页面，同时尽可能选择可信任的信息来源。考虑到部分页面是无法加载的，你也可以多打开一些可能有用的页面而不用担心内容过多。 +`open_url(url: str)`：打开指定的 URL。 + +使用 `【{引用 id}†{引用文本}】` 来引用内容。 + +操作步骤：1. 使用 `search` 来获得信息列表; 2. 使用 `mclick` 来获取指定 ID 页面的内容; 3. 根据获得的内容进行回复。在回复中应当引用信息来源。 + 如果用户提供了 URL，也可以用 `open_url` 直接打开页面。 +如果初次搜索结果没有找到合适的信息，也可以再次使用 `search` 进行搜索。{% elif tool['type'] == 'cogview' %} + +## cogview + +如果用户的请求中包含了对图像的描述，你可以使用 `cogview` 来生成图像并展示给用户。你需要向 `cogview` 发送图像描述，规则： +- 发送给 `cogview` 的消息必须使用英语。用户的中文描述必须完全翻译为英语。 +- 应当尽可能详细地描述图像生成的需求，需求描述约 100 英文单词。 +- 保持用户原始描述的意图。不要虚构内容或者没见过的人物。 +- 如无特殊说明，所在地为中国，持有中国立场并遵循中国社会主义价值观。{% endif %}{% endfor %}{% endif %}{% if item['content'] %}<|{{ item['role'] }}|>{{ item['metadata'] }} +{{ item['content'] }}{% endif %}{% endfor %}{% if add_generation_prompt %}<|assistant|>{% endif %} \ No newline at end of file diff --git a/checkpoint-13796/config.json b/checkpoint-13796/config.json new file mode 100644 index 0000000000000000000000000000000000000000..bc35575d2212a40a1ecaf26ea55c3bdf0649ddfb --- /dev/null +++ b/checkpoint-13796/config.json @@ -0,0 +1,31 @@ +{ + "architectures": [ + "GlmForCausalLM" + ], + "attention_bias": true, + "attention_dropout": 0.0, + "eos_token_id": [ + 151329, + 151336, + 151338 + ], + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 4096, + "initializer_range": 0.02, + "intermediate_size": 13696, + "max_position_embeddings": 131072, + "model_type": "glm", + "num_attention_heads": 32, + "num_hidden_layers": 40, + "num_key_value_heads": 2, + "pad_token_id": 151329, + "partial_rotary_factor": 0.5, + "rms_norm_eps": 1.5625e-07, + "rope_theta": 10000.0, + "tie_word_embeddings": false, + "torch_dtype": "bfloat16", + "transformers_version": "4.53.1", + "use_cache": true, + "vocab_size": 151552 +} diff --git a/checkpoint-13796/generation_config.json b/checkpoint-13796/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..d5a9c67e4e2fe314feb0a9b23144f0efeb644787 --- /dev/null +++ b/checkpoint-13796/generation_config.json @@ -0,0 +1,10 @@ +{ + "_from_model_config": true, + "eos_token_id": [ + 151329, + 151336, + 151338 + ], + "pad_token_id": 151329, + "transformers_version": "4.53.1" +} diff --git a/checkpoint-13796/model-00001-of-00004.safetensors b/checkpoint-13796/model-00001-of-00004.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..93a5d3c0a5be2622b6df62300652b2f894f37c42 --- /dev/null +++ b/checkpoint-13796/model-00001-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e7bd2f01ebc6d2da94caf7aea76f8eab8195e2b117d4603b9699873e797d7fb6 +size 4984133600 diff --git a/checkpoint-13796/model-00002-of-00004.safetensors b/checkpoint-13796/model-00002-of-00004.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..fc98b84b0f96cb86093ead554d72b5cd17669bc8 --- /dev/null +++ b/checkpoint-13796/model-00002-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dc05900c7426c1fe59082ee18756c81947006dbb196af59fe4ff72aa3afeaa3a +size 4895075168 diff --git a/checkpoint-13796/model-00003-of-00004.safetensors b/checkpoint-13796/model-00003-of-00004.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d0678031174e9a6a5a57fb995119c8a184a8ec90 --- /dev/null +++ b/checkpoint-13796/model-00003-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:06a9ec92a94d4ca828edf1470af8ff501f731f436ad1f5d839a2ea81f2de8f0f +size 4895075184 diff --git a/checkpoint-13796/model-00004-of-00004.safetensors b/checkpoint-13796/model-00004-of-00004.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..acef71bd5d941fabce50cb1bfc8410ebe3df7722 --- /dev/null +++ b/checkpoint-13796/model-00004-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a608f3679c3638d21306d4e413ce5441a97a65693fe9083193477b63a0eaba99 +size 4025669744 diff --git a/checkpoint-13796/model.safetensors.index.json b/checkpoint-13796/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..63641162b53de18c620d9bdad367eb87482d7fbf --- /dev/null +++ b/checkpoint-13796/model.safetensors.index.json @@ -0,0 +1,451 @@ +{ + "metadata": { + "total_parameters": 9399951360, + "total_size": 18799902720 + }, + "weight_map": { + "lm_head.weight": "model-00004-of-00004.safetensors", + "model.embed_tokens.weight": "model-00001-of-00004.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.30.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.31.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.32.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.33.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.33.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.33.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.33.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.34.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.34.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.34.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.34.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.35.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.36.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.37.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.38.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.39.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.norm.weight": "model-00004-of-00004.safetensors" + } +} diff --git a/checkpoint-13796/optimizer.pt b/checkpoint-13796/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..669de92c20b7e348b13dc44268207cc771d2b06a --- /dev/null +++ b/checkpoint-13796/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cd92451f74d598366b10266c65464e0fdd68dae59fec74934e191755c7ff7f23 +size 37600193476 diff --git a/checkpoint-13796/rng_state.pth b/checkpoint-13796/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..1feba1a6538e93b94696d3773853dbc8947b0cad --- /dev/null +++ b/checkpoint-13796/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:098b29492211804ab324a36f37466821d948280bb74fce4ba895c03f13ecd878 +size 14645 diff --git a/checkpoint-13796/scheduler.pt b/checkpoint-13796/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..93507bd12ec86d14dbb0ac2315787e1026ef85fa --- /dev/null +++ b/checkpoint-13796/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:067fcefb75f53a4524774245378c4f549aef750a1af171b33d7befad1924ca28 +size 1465 diff --git a/checkpoint-13796/special_tokens_map.json b/checkpoint-13796/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..582da4ad333feaaa0c427b4678a3e48447ea932f --- /dev/null +++ b/checkpoint-13796/special_tokens_map.json @@ -0,0 +1,32 @@ +{ + "additional_special_tokens": [ + "<|endoftext|>", + "[MASK]", + "[gMASK]", + "[sMASK]", + "", + "", + "<|system|>", + "<|user|>", + "<|assistant|>", + "<|observation|>", + "<|begin_of_image|>", + "<|end_of_image|>", + "<|begin_of_video|>", + "<|end_of_video|>" + ], + "eos_token": { + "content": "<|endoftext|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "<|endoftext|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/checkpoint-13796/tokenizer.json b/checkpoint-13796/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..2c24a7b2180f4b088af2671a705217ba4302fee5 --- /dev/null +++ b/checkpoint-13796/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8a7269d6daa6328de533def0082ff9d3a825bb89036cbcc665c324f941f67fbf +size 19967863 diff --git a/checkpoint-13796/tokenizer_config.json b/checkpoint-13796/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..38214260739a871e4c13601f88d904a2ad1fd601 --- /dev/null +++ b/checkpoint-13796/tokenizer_config.json @@ -0,0 +1,145 @@ +{ + "added_tokens_decoder": { + "151329": { + "content": "<|endoftext|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151330": { + "content": "[MASK]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151331": { + "content": "[gMASK]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151332": { + "content": "[sMASK]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151333": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151334": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151335": { + "content": "<|system|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151336": { + "content": "<|user|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151337": { + "content": "<|assistant|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151338": { + "content": "<|observation|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151339": { + "content": "<|begin_of_image|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151340": { + "content": "<|end_of_image|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151341": { + "content": "<|begin_of_video|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151342": { + "content": "<|end_of_video|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "additional_special_tokens": [ + "<|endoftext|>", + "[MASK]", + "[gMASK]", + "[sMASK]", + "", + "", + "<|system|>", + "<|user|>", + "<|assistant|>", + "<|observation|>", + "<|begin_of_image|>", + "<|end_of_image|>", + "<|begin_of_video|>", + "<|end_of_video|>" + ], + "clean_up_tokenization_spaces": false, + "do_lower_case": false, + "eos_token": "<|endoftext|>", + "extra_special_tokens": {}, + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 128000, + "pad_token": "<|endoftext|>", + "padding_side": "left", + "remove_space": false, + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/checkpoint-13796/trainer_state.json b/checkpoint-13796/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..3c08abab0a7825eeb97b672797bbaf5d61be349c --- /dev/null +++ b/checkpoint-13796/trainer_state.json @@ -0,0 +1,9687 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 2.0, + "eval_steps": 500, + "global_step": 13796, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.0014497481062665362, + "grad_norm": 31.75, + "learning_rate": 1.9986952739924617e-05, + "loss": 1.9642, + "step": 10 + }, + { + "epoch": 0.0028994962125330724, + "grad_norm": 10.5, + "learning_rate": 1.99724557842853e-05, + "loss": 2.0457, + "step": 20 + }, + { + "epoch": 0.004349244318799609, + "grad_norm": 8.0, + "learning_rate": 1.9957958828645986e-05, + "loss": 1.7015, + "step": 30 + }, + { + "epoch": 0.005798992425066145, + "grad_norm": 10.4375, + "learning_rate": 1.994346187300667e-05, + "loss": 1.6103, + "step": 40 + }, + { + "epoch": 0.007248740531332681, + "grad_norm": 9.5, + "learning_rate": 1.9928964917367355e-05, + "loss": 1.7592, + "step": 50 + }, + { + "epoch": 0.008698488637599217, + "grad_norm": 5.09375, + "learning_rate": 1.991446796172804e-05, + "loss": 1.114, + "step": 60 + }, + { + "epoch": 0.010148236743865753, + "grad_norm": 5.625, + "learning_rate": 1.9899971006088723e-05, + "loss": 1.1614, + "step": 70 + }, + { + "epoch": 0.01159798485013229, + "grad_norm": 32.5, + "learning_rate": 1.9885474050449408e-05, + "loss": 1.1853, + "step": 80 + }, + { + "epoch": 0.013047732956398826, + "grad_norm": 4.5, + "learning_rate": 1.9870977094810092e-05, + "loss": 1.1573, + "step": 90 + }, + { + "epoch": 0.014497481062665362, + "grad_norm": 6.375, + "learning_rate": 1.9856480139170777e-05, + "loss": 1.183, + "step": 100 + }, + { + "epoch": 0.015947229168931897, + "grad_norm": 4.46875, + "learning_rate": 1.9841983183531458e-05, + "loss": 1.1313, + "step": 110 + }, + { + "epoch": 0.017396977275198434, + "grad_norm": 7.15625, + "learning_rate": 1.9827486227892142e-05, + "loss": 1.0763, + "step": 120 + }, + { + "epoch": 0.01884672538146497, + "grad_norm": 7.0625, + "learning_rate": 1.981298927225283e-05, + "loss": 1.0995, + "step": 130 + }, + { + "epoch": 0.020296473487731507, + "grad_norm": 4.03125, + "learning_rate": 1.9798492316613514e-05, + "loss": 1.0671, + "step": 140 + }, + { + "epoch": 0.02174622159399804, + "grad_norm": 4.5, + "learning_rate": 1.97839953609742e-05, + "loss": 1.0405, + "step": 150 + }, + { + "epoch": 0.02319596970026458, + "grad_norm": 4.15625, + "learning_rate": 1.9769498405334883e-05, + "loss": 1.0388, + "step": 160 + }, + { + "epoch": 0.024645717806531114, + "grad_norm": 4.28125, + "learning_rate": 1.9755001449695564e-05, + "loss": 1.025, + "step": 170 + }, + { + "epoch": 0.02609546591279765, + "grad_norm": 4.25, + "learning_rate": 1.974050449405625e-05, + "loss": 1.0053, + "step": 180 + }, + { + "epoch": 0.027545214019064186, + "grad_norm": 4.15625, + "learning_rate": 1.9726007538416933e-05, + "loss": 0.9725, + "step": 190 + }, + { + "epoch": 0.028994962125330724, + "grad_norm": 3.96875, + "learning_rate": 1.9711510582777617e-05, + "loss": 0.9925, + "step": 200 + }, + { + "epoch": 0.03044471023159726, + "grad_norm": 3.84375, + "learning_rate": 1.9697013627138302e-05, + "loss": 0.9483, + "step": 210 + }, + { + "epoch": 0.03189445833786379, + "grad_norm": 5.71875, + "learning_rate": 1.9682516671498986e-05, + "loss": 0.9845, + "step": 220 + }, + { + "epoch": 0.03334420644413033, + "grad_norm": 302.0, + "learning_rate": 1.966801971585967e-05, + "loss": 5.6501, + "step": 230 + }, + { + "epoch": 0.03479395455039687, + "grad_norm": 40.25, + "learning_rate": 1.9653522760220355e-05, + "loss": 7.2091, + "step": 240 + }, + { + "epoch": 0.03624370265666341, + "grad_norm": 12.625, + "learning_rate": 1.963902580458104e-05, + "loss": 6.3659, + "step": 250 + }, + { + "epoch": 0.03769345076292994, + "grad_norm": 27.625, + "learning_rate": 1.9624528848941724e-05, + "loss": 6.4682, + "step": 260 + }, + { + "epoch": 0.039143198869196476, + "grad_norm": 14.4375, + "learning_rate": 1.961003189330241e-05, + "loss": 6.3851, + "step": 270 + }, + { + "epoch": 0.040592946975463014, + "grad_norm": 12.5625, + "learning_rate": 1.9595534937663093e-05, + "loss": 6.0976, + "step": 280 + }, + { + "epoch": 0.04204269508172955, + "grad_norm": 15.1875, + "learning_rate": 1.9581037982023777e-05, + "loss": 6.1159, + "step": 290 + }, + { + "epoch": 0.04349244318799608, + "grad_norm": 14.5625, + "learning_rate": 1.956654102638446e-05, + "loss": 6.0918, + "step": 300 + }, + { + "epoch": 0.04494219129426262, + "grad_norm": 140.0, + "learning_rate": 1.9552044070745143e-05, + "loss": 6.0004, + "step": 310 + }, + { + "epoch": 0.04639193940052916, + "grad_norm": 79.0, + "learning_rate": 1.953754711510583e-05, + "loss": 6.0565, + "step": 320 + }, + { + "epoch": 0.047841687506795696, + "grad_norm": 65.0, + "learning_rate": 1.9523050159466515e-05, + "loss": 5.7795, + "step": 330 + }, + { + "epoch": 0.04929143561306223, + "grad_norm": 161.0, + "learning_rate": 1.95085532038272e-05, + "loss": 5.4378, + "step": 340 + }, + { + "epoch": 0.050741183719328765, + "grad_norm": 132.0, + "learning_rate": 1.9494056248187884e-05, + "loss": 5.024, + "step": 350 + }, + { + "epoch": 0.0521909318255953, + "grad_norm": 118.5, + "learning_rate": 1.9479559292548568e-05, + "loss": 4.4482, + "step": 360 + }, + { + "epoch": 0.05364067993186184, + "grad_norm": 107.5, + "learning_rate": 1.946506233690925e-05, + "loss": 4.1242, + "step": 370 + }, + { + "epoch": 0.05509042803812837, + "grad_norm": 129.0, + "learning_rate": 1.9450565381269934e-05, + "loss": 3.7731, + "step": 380 + }, + { + "epoch": 0.05654017614439491, + "grad_norm": 117.5, + "learning_rate": 1.9436068425630618e-05, + "loss": 3.6193, + "step": 390 + }, + { + "epoch": 0.05798992425066145, + "grad_norm": 206.0, + "learning_rate": 1.9421571469991302e-05, + "loss": 3.401, + "step": 400 + }, + { + "epoch": 0.059439672356927986, + "grad_norm": 185.0, + "learning_rate": 1.9407074514351987e-05, + "loss": 3.3099, + "step": 410 + }, + { + "epoch": 0.06088942046319452, + "grad_norm": 106.5, + "learning_rate": 1.9392577558712675e-05, + "loss": 3.2412, + "step": 420 + }, + { + "epoch": 0.062339168569461055, + "grad_norm": 163.0, + "learning_rate": 1.9378080603073356e-05, + "loss": 2.9854, + "step": 430 + }, + { + "epoch": 0.06378891667572759, + "grad_norm": 136.0, + "learning_rate": 1.936358364743404e-05, + "loss": 2.9673, + "step": 440 + }, + { + "epoch": 0.06523866478199412, + "grad_norm": 197.0, + "learning_rate": 1.9349086691794724e-05, + "loss": 3.0343, + "step": 450 + }, + { + "epoch": 0.06668841288826066, + "grad_norm": 189.0, + "learning_rate": 1.933458973615541e-05, + "loss": 2.9598, + "step": 460 + }, + { + "epoch": 0.0681381609945272, + "grad_norm": 177.0, + "learning_rate": 1.9320092780516093e-05, + "loss": 2.8282, + "step": 470 + }, + { + "epoch": 0.06958790910079374, + "grad_norm": 108.0, + "learning_rate": 1.9305595824876778e-05, + "loss": 2.8748, + "step": 480 + }, + { + "epoch": 0.07103765720706028, + "grad_norm": 139.0, + "learning_rate": 1.9291098869237462e-05, + "loss": 2.7971, + "step": 490 + }, + { + "epoch": 0.07248740531332681, + "grad_norm": 99.0, + "learning_rate": 1.9276601913598147e-05, + "loss": 2.7033, + "step": 500 + }, + { + "epoch": 0.07393715341959335, + "grad_norm": 187.0, + "learning_rate": 1.9262104957958828e-05, + "loss": 2.747, + "step": 510 + }, + { + "epoch": 0.07538690152585988, + "grad_norm": 114.0, + "learning_rate": 1.9247608002319515e-05, + "loss": 2.6358, + "step": 520 + }, + { + "epoch": 0.07683664963212641, + "grad_norm": 123.5, + "learning_rate": 1.92331110466802e-05, + "loss": 2.6722, + "step": 530 + }, + { + "epoch": 0.07828639773839295, + "grad_norm": 225.0, + "learning_rate": 1.9218614091040884e-05, + "loss": 2.5509, + "step": 540 + }, + { + "epoch": 0.07973614584465949, + "grad_norm": 106.0, + "learning_rate": 1.920411713540157e-05, + "loss": 2.5773, + "step": 550 + }, + { + "epoch": 0.08118589395092603, + "grad_norm": 115.5, + "learning_rate": 1.9189620179762253e-05, + "loss": 2.4187, + "step": 560 + }, + { + "epoch": 0.08263564205719257, + "grad_norm": 129.0, + "learning_rate": 1.9175123224122934e-05, + "loss": 2.4331, + "step": 570 + }, + { + "epoch": 0.0840853901634591, + "grad_norm": 115.5, + "learning_rate": 1.916062626848362e-05, + "loss": 2.388, + "step": 580 + }, + { + "epoch": 0.08553513826972564, + "grad_norm": 141.0, + "learning_rate": 1.9146129312844303e-05, + "loss": 2.4459, + "step": 590 + }, + { + "epoch": 0.08698488637599217, + "grad_norm": 162.0, + "learning_rate": 1.9131632357204987e-05, + "loss": 2.4265, + "step": 600 + }, + { + "epoch": 0.0884346344822587, + "grad_norm": 168.0, + "learning_rate": 1.9117135401565672e-05, + "loss": 2.367, + "step": 610 + }, + { + "epoch": 0.08988438258852524, + "grad_norm": 127.0, + "learning_rate": 1.910263844592636e-05, + "loss": 2.2841, + "step": 620 + }, + { + "epoch": 0.09133413069479178, + "grad_norm": 127.0, + "learning_rate": 1.908814149028704e-05, + "loss": 2.5091, + "step": 630 + }, + { + "epoch": 0.09278387880105832, + "grad_norm": 126.5, + "learning_rate": 1.9073644534647725e-05, + "loss": 2.2486, + "step": 640 + }, + { + "epoch": 0.09423362690732486, + "grad_norm": 100.5, + "learning_rate": 1.905914757900841e-05, + "loss": 2.3366, + "step": 650 + }, + { + "epoch": 0.09568337501359139, + "grad_norm": 135.0, + "learning_rate": 1.9044650623369094e-05, + "loss": 2.3301, + "step": 660 + }, + { + "epoch": 0.09713312311985793, + "grad_norm": 136.0, + "learning_rate": 1.9030153667729778e-05, + "loss": 2.3708, + "step": 670 + }, + { + "epoch": 0.09858287122612445, + "grad_norm": 118.0, + "learning_rate": 1.9015656712090463e-05, + "loss": 2.2925, + "step": 680 + }, + { + "epoch": 0.10003261933239099, + "grad_norm": 131.0, + "learning_rate": 1.9001159756451147e-05, + "loss": 2.2805, + "step": 690 + }, + { + "epoch": 0.10148236743865753, + "grad_norm": 93.0, + "learning_rate": 1.898666280081183e-05, + "loss": 2.2899, + "step": 700 + }, + { + "epoch": 0.10293211554492407, + "grad_norm": 135.0, + "learning_rate": 1.8972165845172516e-05, + "loss": 2.0617, + "step": 710 + }, + { + "epoch": 0.1043818636511906, + "grad_norm": 136.0, + "learning_rate": 1.89576688895332e-05, + "loss": 2.2515, + "step": 720 + }, + { + "epoch": 0.10583161175745714, + "grad_norm": 178.0, + "learning_rate": 1.8943171933893885e-05, + "loss": 2.1664, + "step": 730 + }, + { + "epoch": 0.10728135986372368, + "grad_norm": 158.0, + "learning_rate": 1.892867497825457e-05, + "loss": 2.1863, + "step": 740 + }, + { + "epoch": 0.10873110796999022, + "grad_norm": 133.0, + "learning_rate": 1.8914178022615254e-05, + "loss": 2.169, + "step": 750 + }, + { + "epoch": 0.11018085607625674, + "grad_norm": 111.5, + "learning_rate": 1.8899681066975938e-05, + "loss": 2.1845, + "step": 760 + }, + { + "epoch": 0.11163060418252328, + "grad_norm": 101.5, + "learning_rate": 1.8885184111336622e-05, + "loss": 2.1846, + "step": 770 + }, + { + "epoch": 0.11308035228878982, + "grad_norm": 112.0, + "learning_rate": 1.8870687155697303e-05, + "loss": 2.1391, + "step": 780 + }, + { + "epoch": 0.11453010039505636, + "grad_norm": 135.0, + "learning_rate": 1.8856190200057988e-05, + "loss": 2.1059, + "step": 790 + }, + { + "epoch": 0.1159798485013229, + "grad_norm": 256.0, + "learning_rate": 1.8841693244418672e-05, + "loss": 2.2222, + "step": 800 + }, + { + "epoch": 0.11742959660758943, + "grad_norm": 174.0, + "learning_rate": 1.8827196288779357e-05, + "loss": 2.0608, + "step": 810 + }, + { + "epoch": 0.11887934471385597, + "grad_norm": 140.0, + "learning_rate": 1.8812699333140045e-05, + "loss": 2.0371, + "step": 820 + }, + { + "epoch": 0.12032909282012251, + "grad_norm": 159.0, + "learning_rate": 1.8798202377500726e-05, + "loss": 2.045, + "step": 830 + }, + { + "epoch": 0.12177884092638903, + "grad_norm": 103.5, + "learning_rate": 1.878370542186141e-05, + "loss": 1.9997, + "step": 840 + }, + { + "epoch": 0.12322858903265557, + "grad_norm": 121.0, + "learning_rate": 1.8769208466222094e-05, + "loss": 2.0391, + "step": 850 + }, + { + "epoch": 0.12467833713892211, + "grad_norm": 122.5, + "learning_rate": 1.875471151058278e-05, + "loss": 2.062, + "step": 860 + }, + { + "epoch": 0.12612808524518865, + "grad_norm": 164.0, + "learning_rate": 1.8740214554943463e-05, + "loss": 2.1222, + "step": 870 + }, + { + "epoch": 0.12757783335145517, + "grad_norm": 110.5, + "learning_rate": 1.8725717599304148e-05, + "loss": 2.0443, + "step": 880 + }, + { + "epoch": 0.12902758145772172, + "grad_norm": 97.0, + "learning_rate": 1.8711220643664832e-05, + "loss": 1.9237, + "step": 890 + }, + { + "epoch": 0.13047732956398825, + "grad_norm": 103.5, + "learning_rate": 1.8696723688025516e-05, + "loss": 1.9752, + "step": 900 + }, + { + "epoch": 0.1319270776702548, + "grad_norm": 109.5, + "learning_rate": 1.86822267323862e-05, + "loss": 2.0141, + "step": 910 + }, + { + "epoch": 0.13337682577652132, + "grad_norm": 132.0, + "learning_rate": 1.8667729776746885e-05, + "loss": 1.9615, + "step": 920 + }, + { + "epoch": 0.13482657388278788, + "grad_norm": 102.0, + "learning_rate": 1.865323282110757e-05, + "loss": 1.9503, + "step": 930 + }, + { + "epoch": 0.1362763219890544, + "grad_norm": 117.5, + "learning_rate": 1.8638735865468254e-05, + "loss": 1.8662, + "step": 940 + }, + { + "epoch": 0.13772607009532095, + "grad_norm": 133.0, + "learning_rate": 1.862423890982894e-05, + "loss": 1.965, + "step": 950 + }, + { + "epoch": 0.13917581820158748, + "grad_norm": 114.0, + "learning_rate": 1.8609741954189623e-05, + "loss": 1.8814, + "step": 960 + }, + { + "epoch": 0.140625566307854, + "grad_norm": 140.0, + "learning_rate": 1.8595244998550307e-05, + "loss": 1.9293, + "step": 970 + }, + { + "epoch": 0.14207531441412055, + "grad_norm": 121.5, + "learning_rate": 1.858074804291099e-05, + "loss": 1.9772, + "step": 980 + }, + { + "epoch": 0.14352506252038708, + "grad_norm": 120.0, + "learning_rate": 1.8566251087271673e-05, + "loss": 1.9769, + "step": 990 + }, + { + "epoch": 0.14497481062665363, + "grad_norm": 70.5, + "learning_rate": 1.8551754131632357e-05, + "loss": 1.9106, + "step": 1000 + }, + { + "epoch": 0.14642455873292015, + "grad_norm": 132.0, + "learning_rate": 1.853725717599304e-05, + "loss": 1.9397, + "step": 1010 + }, + { + "epoch": 0.1478743068391867, + "grad_norm": 120.5, + "learning_rate": 1.852276022035373e-05, + "loss": 1.8997, + "step": 1020 + }, + { + "epoch": 0.14932405494545323, + "grad_norm": 110.5, + "learning_rate": 1.8508263264714414e-05, + "loss": 1.8802, + "step": 1030 + }, + { + "epoch": 0.15077380305171975, + "grad_norm": 130.0, + "learning_rate": 1.8493766309075095e-05, + "loss": 1.9104, + "step": 1040 + }, + { + "epoch": 0.1522235511579863, + "grad_norm": 127.0, + "learning_rate": 1.847926935343578e-05, + "loss": 1.8504, + "step": 1050 + }, + { + "epoch": 0.15367329926425283, + "grad_norm": 103.0, + "learning_rate": 1.8464772397796464e-05, + "loss": 1.8127, + "step": 1060 + }, + { + "epoch": 0.15512304737051938, + "grad_norm": 113.0, + "learning_rate": 1.8450275442157148e-05, + "loss": 1.8785, + "step": 1070 + }, + { + "epoch": 0.1565727954767859, + "grad_norm": 98.5, + "learning_rate": 1.8435778486517833e-05, + "loss": 1.7915, + "step": 1080 + }, + { + "epoch": 0.15802254358305245, + "grad_norm": 428.0, + "learning_rate": 1.8421281530878517e-05, + "loss": 3.0407, + "step": 1090 + }, + { + "epoch": 0.15947229168931898, + "grad_norm": 362.0, + "learning_rate": 1.84067845752392e-05, + "loss": 2.29, + "step": 1100 + }, + { + "epoch": 0.16092203979558553, + "grad_norm": 1312.0, + "learning_rate": 1.8392287619599886e-05, + "loss": 2.7752, + "step": 1110 + }, + { + "epoch": 0.16237178790185205, + "grad_norm": 576.0, + "learning_rate": 1.837779066396057e-05, + "loss": 2.8033, + "step": 1120 + }, + { + "epoch": 0.16382153600811858, + "grad_norm": 153.0, + "learning_rate": 1.8363293708321255e-05, + "loss": 2.5284, + "step": 1130 + }, + { + "epoch": 0.16527128411438513, + "grad_norm": 454.0, + "learning_rate": 1.834879675268194e-05, + "loss": 3.7134, + "step": 1140 + }, + { + "epoch": 0.16672103222065165, + "grad_norm": 2008.0, + "learning_rate": 1.8334299797042623e-05, + "loss": 2.2115, + "step": 1150 + }, + { + "epoch": 0.1681707803269182, + "grad_norm": 298.0, + "learning_rate": 1.8319802841403308e-05, + "loss": 1.7636, + "step": 1160 + }, + { + "epoch": 0.16962052843318473, + "grad_norm": 6.65625, + "learning_rate": 1.8305305885763992e-05, + "loss": 1.2321, + "step": 1170 + }, + { + "epoch": 0.17107027653945128, + "grad_norm": 5.15625, + "learning_rate": 1.8290808930124673e-05, + "loss": 1.0327, + "step": 1180 + }, + { + "epoch": 0.1725200246457178, + "grad_norm": 9.5, + "learning_rate": 1.8276311974485358e-05, + "loss": 1.0053, + "step": 1190 + }, + { + "epoch": 0.17396977275198433, + "grad_norm": 10.125, + "learning_rate": 1.8261815018846042e-05, + "loss": 1.0167, + "step": 1200 + }, + { + "epoch": 0.17541952085825088, + "grad_norm": 4.1875, + "learning_rate": 1.8247318063206727e-05, + "loss": 1.0274, + "step": 1210 + }, + { + "epoch": 0.1768692689645174, + "grad_norm": 4.625, + "learning_rate": 1.8232821107567414e-05, + "loss": 1.0167, + "step": 1220 + }, + { + "epoch": 0.17831901707078396, + "grad_norm": 3.984375, + "learning_rate": 1.82183241519281e-05, + "loss": 1.0135, + "step": 1230 + }, + { + "epoch": 0.17976876517705048, + "grad_norm": 13.1875, + "learning_rate": 1.820382719628878e-05, + "loss": 1.0732, + "step": 1240 + }, + { + "epoch": 0.18121851328331703, + "grad_norm": 4.28125, + "learning_rate": 1.8189330240649464e-05, + "loss": 1.0478, + "step": 1250 + }, + { + "epoch": 0.18266826138958356, + "grad_norm": 4.375, + "learning_rate": 1.817483328501015e-05, + "loss": 0.9829, + "step": 1260 + }, + { + "epoch": 0.18411800949585008, + "grad_norm": 3.96875, + "learning_rate": 1.8160336329370833e-05, + "loss": 1.0159, + "step": 1270 + }, + { + "epoch": 0.18556775760211663, + "grad_norm": 3.859375, + "learning_rate": 1.8145839373731518e-05, + "loss": 1.0163, + "step": 1280 + }, + { + "epoch": 0.18701750570838316, + "grad_norm": 3.765625, + "learning_rate": 1.8131342418092202e-05, + "loss": 0.954, + "step": 1290 + }, + { + "epoch": 0.1884672538146497, + "grad_norm": 4.65625, + "learning_rate": 1.8116845462452886e-05, + "loss": 0.9436, + "step": 1300 + }, + { + "epoch": 0.18991700192091623, + "grad_norm": 3.953125, + "learning_rate": 1.810234850681357e-05, + "loss": 0.9598, + "step": 1310 + }, + { + "epoch": 0.19136675002718279, + "grad_norm": 4.375, + "learning_rate": 1.8087851551174255e-05, + "loss": 0.9772, + "step": 1320 + }, + { + "epoch": 0.1928164981334493, + "grad_norm": 4.25, + "learning_rate": 1.807335459553494e-05, + "loss": 0.9666, + "step": 1330 + }, + { + "epoch": 0.19426624623971586, + "grad_norm": 4.4375, + "learning_rate": 1.8058857639895624e-05, + "loss": 0.9915, + "step": 1340 + }, + { + "epoch": 0.19571599434598239, + "grad_norm": 3.65625, + "learning_rate": 1.804436068425631e-05, + "loss": 0.9198, + "step": 1350 + }, + { + "epoch": 0.1971657424522489, + "grad_norm": 4.28125, + "learning_rate": 1.8029863728616993e-05, + "loss": 0.9415, + "step": 1360 + }, + { + "epoch": 0.19861549055851546, + "grad_norm": 4.5, + "learning_rate": 1.8015366772977677e-05, + "loss": 0.8826, + "step": 1370 + }, + { + "epoch": 0.20006523866478199, + "grad_norm": 4.375, + "learning_rate": 1.8000869817338362e-05, + "loss": 0.9681, + "step": 1380 + }, + { + "epoch": 0.20151498677104854, + "grad_norm": 3.890625, + "learning_rate": 1.7986372861699043e-05, + "loss": 0.9526, + "step": 1390 + }, + { + "epoch": 0.20296473487731506, + "grad_norm": 3.625, + "learning_rate": 1.7971875906059727e-05, + "loss": 0.9296, + "step": 1400 + }, + { + "epoch": 0.2044144829835816, + "grad_norm": 3.5, + "learning_rate": 1.795737895042041e-05, + "loss": 0.9136, + "step": 1410 + }, + { + "epoch": 0.20586423108984814, + "grad_norm": 3.578125, + "learning_rate": 1.79428819947811e-05, + "loss": 0.9171, + "step": 1420 + }, + { + "epoch": 0.20731397919611466, + "grad_norm": 19.75, + "learning_rate": 1.7928385039141784e-05, + "loss": 0.8676, + "step": 1430 + }, + { + "epoch": 0.2087637273023812, + "grad_norm": 3.34375, + "learning_rate": 1.7913888083502465e-05, + "loss": 0.8419, + "step": 1440 + }, + { + "epoch": 0.21021347540864774, + "grad_norm": 4.96875, + "learning_rate": 1.789939112786315e-05, + "loss": 0.8553, + "step": 1450 + }, + { + "epoch": 0.2116632235149143, + "grad_norm": 3.703125, + "learning_rate": 1.7884894172223834e-05, + "loss": 0.8858, + "step": 1460 + }, + { + "epoch": 0.2131129716211808, + "grad_norm": 9.875, + "learning_rate": 1.7870397216584518e-05, + "loss": 0.8399, + "step": 1470 + }, + { + "epoch": 0.21456271972744737, + "grad_norm": 2.96875, + "learning_rate": 1.7855900260945202e-05, + "loss": 0.8398, + "step": 1480 + }, + { + "epoch": 0.2160124678337139, + "grad_norm": 3.515625, + "learning_rate": 1.7841403305305887e-05, + "loss": 0.8399, + "step": 1490 + }, + { + "epoch": 0.21746221593998044, + "grad_norm": 3.15625, + "learning_rate": 1.782690634966657e-05, + "loss": 0.7819, + "step": 1500 + }, + { + "epoch": 0.21891196404624697, + "grad_norm": 3.40625, + "learning_rate": 1.7812409394027256e-05, + "loss": 0.845, + "step": 1510 + }, + { + "epoch": 0.2203617121525135, + "grad_norm": 3.015625, + "learning_rate": 1.779791243838794e-05, + "loss": 0.8055, + "step": 1520 + }, + { + "epoch": 0.22181146025878004, + "grad_norm": 3.140625, + "learning_rate": 1.7783415482748625e-05, + "loss": 0.8237, + "step": 1530 + }, + { + "epoch": 0.22326120836504657, + "grad_norm": 3.15625, + "learning_rate": 1.776891852710931e-05, + "loss": 0.8217, + "step": 1540 + }, + { + "epoch": 0.22471095647131312, + "grad_norm": 2.859375, + "learning_rate": 1.7754421571469993e-05, + "loss": 0.815, + "step": 1550 + }, + { + "epoch": 0.22616070457757964, + "grad_norm": 2.953125, + "learning_rate": 1.7739924615830678e-05, + "loss": 0.8116, + "step": 1560 + }, + { + "epoch": 0.2276104526838462, + "grad_norm": 2.859375, + "learning_rate": 1.7725427660191362e-05, + "loss": 0.8284, + "step": 1570 + }, + { + "epoch": 0.22906020079011272, + "grad_norm": 3.65625, + "learning_rate": 1.7710930704552047e-05, + "loss": 0.8, + "step": 1580 + }, + { + "epoch": 0.23050994889637924, + "grad_norm": 2.875, + "learning_rate": 1.7696433748912728e-05, + "loss": 0.7552, + "step": 1590 + }, + { + "epoch": 0.2319596970026458, + "grad_norm": 2.875, + "learning_rate": 1.7681936793273412e-05, + "loss": 0.7149, + "step": 1600 + }, + { + "epoch": 0.23340944510891232, + "grad_norm": 3.171875, + "learning_rate": 1.7667439837634097e-05, + "loss": 0.7254, + "step": 1610 + }, + { + "epoch": 0.23485919321517887, + "grad_norm": 3.25, + "learning_rate": 1.7652942881994784e-05, + "loss": 0.7839, + "step": 1620 + }, + { + "epoch": 0.2363089413214454, + "grad_norm": 3.265625, + "learning_rate": 1.763844592635547e-05, + "loss": 0.7774, + "step": 1630 + }, + { + "epoch": 0.23775868942771194, + "grad_norm": 3.1875, + "learning_rate": 1.7623948970716153e-05, + "loss": 0.771, + "step": 1640 + }, + { + "epoch": 0.23920843753397847, + "grad_norm": 2.859375, + "learning_rate": 1.7609452015076834e-05, + "loss": 0.748, + "step": 1650 + }, + { + "epoch": 0.24065818564024502, + "grad_norm": 3.09375, + "learning_rate": 1.759495505943752e-05, + "loss": 0.7521, + "step": 1660 + }, + { + "epoch": 0.24210793374651154, + "grad_norm": 3.140625, + "learning_rate": 1.7580458103798203e-05, + "loss": 0.7102, + "step": 1670 + }, + { + "epoch": 0.24355768185277807, + "grad_norm": 2.984375, + "learning_rate": 1.7565961148158887e-05, + "loss": 0.7081, + "step": 1680 + }, + { + "epoch": 0.24500742995904462, + "grad_norm": 3.28125, + "learning_rate": 1.7551464192519572e-05, + "loss": 0.7104, + "step": 1690 + }, + { + "epoch": 0.24645717806531114, + "grad_norm": 3.28125, + "learning_rate": 1.7536967236880256e-05, + "loss": 0.6859, + "step": 1700 + }, + { + "epoch": 0.2479069261715777, + "grad_norm": 3.046875, + "learning_rate": 1.752247028124094e-05, + "loss": 0.6955, + "step": 1710 + }, + { + "epoch": 0.24935667427784422, + "grad_norm": 3.546875, + "learning_rate": 1.7507973325601625e-05, + "loss": 0.7151, + "step": 1720 + }, + { + "epoch": 0.25080642238411077, + "grad_norm": 2.859375, + "learning_rate": 1.749347636996231e-05, + "loss": 0.74, + "step": 1730 + }, + { + "epoch": 0.2522561704903773, + "grad_norm": 2.546875, + "learning_rate": 1.7478979414322994e-05, + "loss": 0.6606, + "step": 1740 + }, + { + "epoch": 0.2537059185966438, + "grad_norm": 2.765625, + "learning_rate": 1.746448245868368e-05, + "loss": 0.6891, + "step": 1750 + }, + { + "epoch": 0.25515566670291034, + "grad_norm": 2.953125, + "learning_rate": 1.7449985503044363e-05, + "loss": 0.7079, + "step": 1760 + }, + { + "epoch": 0.2566054148091769, + "grad_norm": 3.03125, + "learning_rate": 1.7435488547405047e-05, + "loss": 0.7503, + "step": 1770 + }, + { + "epoch": 0.25805516291544345, + "grad_norm": 3.265625, + "learning_rate": 1.742099159176573e-05, + "loss": 0.6945, + "step": 1780 + }, + { + "epoch": 0.25950491102170997, + "grad_norm": 2.84375, + "learning_rate": 1.7406494636126413e-05, + "loss": 0.6704, + "step": 1790 + }, + { + "epoch": 0.2609546591279765, + "grad_norm": 3.0, + "learning_rate": 1.7391997680487097e-05, + "loss": 0.6345, + "step": 1800 + }, + { + "epoch": 0.2624044072342431, + "grad_norm": 3.0, + "learning_rate": 1.7377500724847785e-05, + "loss": 0.6334, + "step": 1810 + }, + { + "epoch": 0.2638541553405096, + "grad_norm": 2.96875, + "learning_rate": 1.736300376920847e-05, + "loss": 0.6747, + "step": 1820 + }, + { + "epoch": 0.2653039034467761, + "grad_norm": 3.015625, + "learning_rate": 1.7348506813569154e-05, + "loss": 0.6682, + "step": 1830 + }, + { + "epoch": 0.26675365155304265, + "grad_norm": 2.96875, + "learning_rate": 1.7334009857929838e-05, + "loss": 0.6314, + "step": 1840 + }, + { + "epoch": 0.26820339965930917, + "grad_norm": 3.1875, + "learning_rate": 1.731951290229052e-05, + "loss": 0.6588, + "step": 1850 + }, + { + "epoch": 0.26965314776557575, + "grad_norm": 2.59375, + "learning_rate": 1.7305015946651204e-05, + "loss": 0.663, + "step": 1860 + }, + { + "epoch": 0.2711028958718423, + "grad_norm": 2.921875, + "learning_rate": 1.7290518991011888e-05, + "loss": 0.6683, + "step": 1870 + }, + { + "epoch": 0.2725526439781088, + "grad_norm": 2.875, + "learning_rate": 1.7276022035372572e-05, + "loss": 0.6402, + "step": 1880 + }, + { + "epoch": 0.2740023920843753, + "grad_norm": 2.75, + "learning_rate": 1.7261525079733257e-05, + "loss": 0.6216, + "step": 1890 + }, + { + "epoch": 0.2754521401906419, + "grad_norm": 2.9375, + "learning_rate": 1.724702812409394e-05, + "loss": 0.6139, + "step": 1900 + }, + { + "epoch": 0.2769018882969084, + "grad_norm": 3.328125, + "learning_rate": 1.7232531168454626e-05, + "loss": 0.6549, + "step": 1910 + }, + { + "epoch": 0.27835163640317495, + "grad_norm": 2.96875, + "learning_rate": 1.721803421281531e-05, + "loss": 0.5732, + "step": 1920 + }, + { + "epoch": 0.2798013845094415, + "grad_norm": 2.640625, + "learning_rate": 1.7203537257175994e-05, + "loss": 0.6259, + "step": 1930 + }, + { + "epoch": 0.281251132615708, + "grad_norm": 2.734375, + "learning_rate": 1.718904030153668e-05, + "loss": 0.63, + "step": 1940 + }, + { + "epoch": 0.2827008807219746, + "grad_norm": 2.78125, + "learning_rate": 1.7174543345897363e-05, + "loss": 0.6002, + "step": 1950 + }, + { + "epoch": 0.2841506288282411, + "grad_norm": 2.96875, + "learning_rate": 1.7160046390258048e-05, + "loss": 0.6192, + "step": 1960 + }, + { + "epoch": 0.2856003769345076, + "grad_norm": 3.125, + "learning_rate": 1.7145549434618732e-05, + "loss": 0.5734, + "step": 1970 + }, + { + "epoch": 0.28705012504077415, + "grad_norm": 2.953125, + "learning_rate": 1.7131052478979417e-05, + "loss": 0.5924, + "step": 1980 + }, + { + "epoch": 0.2884998731470407, + "grad_norm": 2.859375, + "learning_rate": 1.71165555233401e-05, + "loss": 0.5542, + "step": 1990 + }, + { + "epoch": 0.28994962125330725, + "grad_norm": 2.703125, + "learning_rate": 1.7102058567700782e-05, + "loss": 0.5508, + "step": 2000 + }, + { + "epoch": 0.2913993693595738, + "grad_norm": 2.5, + "learning_rate": 1.708756161206147e-05, + "loss": 0.5542, + "step": 2010 + }, + { + "epoch": 0.2928491174658403, + "grad_norm": 2.90625, + "learning_rate": 1.7073064656422154e-05, + "loss": 0.564, + "step": 2020 + }, + { + "epoch": 0.2942988655721068, + "grad_norm": 2.84375, + "learning_rate": 1.705856770078284e-05, + "loss": 0.5562, + "step": 2030 + }, + { + "epoch": 0.2957486136783734, + "grad_norm": 2.953125, + "learning_rate": 1.7044070745143523e-05, + "loss": 0.608, + "step": 2040 + }, + { + "epoch": 0.29719836178463993, + "grad_norm": 2.421875, + "learning_rate": 1.7029573789504204e-05, + "loss": 0.559, + "step": 2050 + }, + { + "epoch": 0.29864810989090645, + "grad_norm": 2.828125, + "learning_rate": 1.701507683386489e-05, + "loss": 0.594, + "step": 2060 + }, + { + "epoch": 0.300097857997173, + "grad_norm": 2.703125, + "learning_rate": 1.7000579878225573e-05, + "loss": 0.5499, + "step": 2070 + }, + { + "epoch": 0.3015476061034395, + "grad_norm": 2.671875, + "learning_rate": 1.6986082922586257e-05, + "loss": 0.5656, + "step": 2080 + }, + { + "epoch": 0.3029973542097061, + "grad_norm": 2.625, + "learning_rate": 1.6971585966946942e-05, + "loss": 0.552, + "step": 2090 + }, + { + "epoch": 0.3044471023159726, + "grad_norm": 3.046875, + "learning_rate": 1.6957089011307626e-05, + "loss": 0.606, + "step": 2100 + }, + { + "epoch": 0.30589685042223913, + "grad_norm": 2.78125, + "learning_rate": 1.694259205566831e-05, + "loss": 0.5693, + "step": 2110 + }, + { + "epoch": 0.30734659852850565, + "grad_norm": 3.1875, + "learning_rate": 1.6928095100028995e-05, + "loss": 0.5844, + "step": 2120 + }, + { + "epoch": 0.30879634663477223, + "grad_norm": 2.71875, + "learning_rate": 1.691359814438968e-05, + "loss": 0.5218, + "step": 2130 + }, + { + "epoch": 0.31024609474103876, + "grad_norm": 2.765625, + "learning_rate": 1.6899101188750364e-05, + "loss": 0.5425, + "step": 2140 + }, + { + "epoch": 0.3116958428473053, + "grad_norm": 2.796875, + "learning_rate": 1.6884604233111048e-05, + "loss": 0.5616, + "step": 2150 + }, + { + "epoch": 0.3131455909535718, + "grad_norm": 2.828125, + "learning_rate": 1.6870107277471733e-05, + "loss": 0.557, + "step": 2160 + }, + { + "epoch": 0.31459533905983833, + "grad_norm": 2.71875, + "learning_rate": 1.6855610321832417e-05, + "loss": 0.5053, + "step": 2170 + }, + { + "epoch": 0.3160450871661049, + "grad_norm": 3.078125, + "learning_rate": 1.68411133661931e-05, + "loss": 0.5129, + "step": 2180 + }, + { + "epoch": 0.31749483527237143, + "grad_norm": 2.546875, + "learning_rate": 1.6826616410553786e-05, + "loss": 0.5499, + "step": 2190 + }, + { + "epoch": 0.31894458337863796, + "grad_norm": 2.4375, + "learning_rate": 1.6812119454914467e-05, + "loss": 0.5033, + "step": 2200 + }, + { + "epoch": 0.3203943314849045, + "grad_norm": 2.84375, + "learning_rate": 1.6797622499275155e-05, + "loss": 0.5304, + "step": 2210 + }, + { + "epoch": 0.32184407959117106, + "grad_norm": 2.921875, + "learning_rate": 1.678312554363584e-05, + "loss": 0.5024, + "step": 2220 + }, + { + "epoch": 0.3232938276974376, + "grad_norm": 2.5625, + "learning_rate": 1.6768628587996524e-05, + "loss": 0.497, + "step": 2230 + }, + { + "epoch": 0.3247435758037041, + "grad_norm": 2.640625, + "learning_rate": 1.6754131632357208e-05, + "loss": 0.5155, + "step": 2240 + }, + { + "epoch": 0.32619332390997063, + "grad_norm": 2.671875, + "learning_rate": 1.6739634676717892e-05, + "loss": 0.5126, + "step": 2250 + }, + { + "epoch": 0.32764307201623716, + "grad_norm": 2.78125, + "learning_rate": 1.6725137721078573e-05, + "loss": 0.5221, + "step": 2260 + }, + { + "epoch": 0.32909282012250374, + "grad_norm": 2.765625, + "learning_rate": 1.6710640765439258e-05, + "loss": 0.5153, + "step": 2270 + }, + { + "epoch": 0.33054256822877026, + "grad_norm": 2.9375, + "learning_rate": 1.6696143809799942e-05, + "loss": 0.5342, + "step": 2280 + }, + { + "epoch": 0.3319923163350368, + "grad_norm": 2.625, + "learning_rate": 1.6681646854160627e-05, + "loss": 0.4855, + "step": 2290 + }, + { + "epoch": 0.3334420644413033, + "grad_norm": 2.78125, + "learning_rate": 1.666714989852131e-05, + "loss": 0.4692, + "step": 2300 + }, + { + "epoch": 0.33489181254756983, + "grad_norm": 2.515625, + "learning_rate": 1.6652652942881996e-05, + "loss": 0.4713, + "step": 2310 + }, + { + "epoch": 0.3363415606538364, + "grad_norm": 2.640625, + "learning_rate": 1.663815598724268e-05, + "loss": 0.4787, + "step": 2320 + }, + { + "epoch": 0.33779130876010294, + "grad_norm": 2.734375, + "learning_rate": 1.6623659031603364e-05, + "loss": 0.4909, + "step": 2330 + }, + { + "epoch": 0.33924105686636946, + "grad_norm": 2.734375, + "learning_rate": 1.660916207596405e-05, + "loss": 0.4572, + "step": 2340 + }, + { + "epoch": 0.340690804972636, + "grad_norm": 2.359375, + "learning_rate": 1.6594665120324733e-05, + "loss": 0.4744, + "step": 2350 + }, + { + "epoch": 0.34214055307890257, + "grad_norm": 2.734375, + "learning_rate": 1.6580168164685418e-05, + "loss": 0.4718, + "step": 2360 + }, + { + "epoch": 0.3435903011851691, + "grad_norm": 2.734375, + "learning_rate": 1.6565671209046102e-05, + "loss": 0.4431, + "step": 2370 + }, + { + "epoch": 0.3450400492914356, + "grad_norm": 2.796875, + "learning_rate": 1.6551174253406786e-05, + "loss": 0.4819, + "step": 2380 + }, + { + "epoch": 0.34648979739770214, + "grad_norm": 3.09375, + "learning_rate": 1.653667729776747e-05, + "loss": 0.4923, + "step": 2390 + }, + { + "epoch": 0.34793954550396866, + "grad_norm": 2.953125, + "learning_rate": 1.6522180342128152e-05, + "loss": 0.4772, + "step": 2400 + }, + { + "epoch": 0.34938929361023524, + "grad_norm": 2.515625, + "learning_rate": 1.650768338648884e-05, + "loss": 0.4737, + "step": 2410 + }, + { + "epoch": 0.35083904171650176, + "grad_norm": 3.3125, + "learning_rate": 1.6493186430849524e-05, + "loss": 0.4902, + "step": 2420 + }, + { + "epoch": 0.3522887898227683, + "grad_norm": 3.515625, + "learning_rate": 1.647868947521021e-05, + "loss": 0.4489, + "step": 2430 + }, + { + "epoch": 0.3537385379290348, + "grad_norm": 2.484375, + "learning_rate": 1.6464192519570893e-05, + "loss": 0.4554, + "step": 2440 + }, + { + "epoch": 0.3551882860353014, + "grad_norm": 2.6875, + "learning_rate": 1.6449695563931577e-05, + "loss": 0.4957, + "step": 2450 + }, + { + "epoch": 0.3566380341415679, + "grad_norm": 2.765625, + "learning_rate": 1.643519860829226e-05, + "loss": 0.439, + "step": 2460 + }, + { + "epoch": 0.35808778224783444, + "grad_norm": 2.703125, + "learning_rate": 1.6420701652652943e-05, + "loss": 0.4508, + "step": 2470 + }, + { + "epoch": 0.35953753035410096, + "grad_norm": 2.625, + "learning_rate": 1.6406204697013627e-05, + "loss": 0.4258, + "step": 2480 + }, + { + "epoch": 0.3609872784603675, + "grad_norm": 2.71875, + "learning_rate": 1.639170774137431e-05, + "loss": 0.4699, + "step": 2490 + }, + { + "epoch": 0.36243702656663407, + "grad_norm": 2.578125, + "learning_rate": 1.6377210785734996e-05, + "loss": 0.432, + "step": 2500 + }, + { + "epoch": 0.3638867746729006, + "grad_norm": 2.3125, + "learning_rate": 1.6362713830095684e-05, + "loss": 0.4304, + "step": 2510 + }, + { + "epoch": 0.3653365227791671, + "grad_norm": 2.78125, + "learning_rate": 1.6348216874456365e-05, + "loss": 0.4805, + "step": 2520 + }, + { + "epoch": 0.36678627088543364, + "grad_norm": 2.453125, + "learning_rate": 1.633371991881705e-05, + "loss": 0.4604, + "step": 2530 + }, + { + "epoch": 0.36823601899170016, + "grad_norm": 2.265625, + "learning_rate": 1.6319222963177734e-05, + "loss": 0.4803, + "step": 2540 + }, + { + "epoch": 0.36968576709796674, + "grad_norm": 2.65625, + "learning_rate": 1.6304726007538418e-05, + "loss": 0.4259, + "step": 2550 + }, + { + "epoch": 0.37113551520423327, + "grad_norm": 2.890625, + "learning_rate": 1.6290229051899103e-05, + "loss": 0.4283, + "step": 2560 + }, + { + "epoch": 0.3725852633104998, + "grad_norm": 2.65625, + "learning_rate": 1.6275732096259787e-05, + "loss": 0.4135, + "step": 2570 + }, + { + "epoch": 0.3740350114167663, + "grad_norm": 2.625, + "learning_rate": 1.626123514062047e-05, + "loss": 0.4054, + "step": 2580 + }, + { + "epoch": 0.3754847595230329, + "grad_norm": 2.890625, + "learning_rate": 1.6246738184981156e-05, + "loss": 0.3739, + "step": 2590 + }, + { + "epoch": 0.3769345076292994, + "grad_norm": 2.453125, + "learning_rate": 1.623224122934184e-05, + "loss": 0.417, + "step": 2600 + }, + { + "epoch": 0.37838425573556594, + "grad_norm": 2.671875, + "learning_rate": 1.6217744273702525e-05, + "loss": 0.3945, + "step": 2610 + }, + { + "epoch": 0.37983400384183247, + "grad_norm": 2.515625, + "learning_rate": 1.620324731806321e-05, + "loss": 0.4265, + "step": 2620 + }, + { + "epoch": 0.381283751948099, + "grad_norm": 2.34375, + "learning_rate": 1.6188750362423893e-05, + "loss": 0.4437, + "step": 2630 + }, + { + "epoch": 0.38273350005436557, + "grad_norm": 2.8125, + "learning_rate": 1.6174253406784578e-05, + "loss": 0.4536, + "step": 2640 + }, + { + "epoch": 0.3841832481606321, + "grad_norm": 2.96875, + "learning_rate": 1.6159756451145262e-05, + "loss": 0.4168, + "step": 2650 + }, + { + "epoch": 0.3856329962668986, + "grad_norm": 2.578125, + "learning_rate": 1.6145259495505943e-05, + "loss": 0.4287, + "step": 2660 + }, + { + "epoch": 0.38708274437316514, + "grad_norm": 2.9375, + "learning_rate": 1.6130762539866628e-05, + "loss": 0.433, + "step": 2670 + }, + { + "epoch": 0.3885324924794317, + "grad_norm": 2.703125, + "learning_rate": 1.6116265584227312e-05, + "loss": 0.4092, + "step": 2680 + }, + { + "epoch": 0.38998224058569825, + "grad_norm": 2.390625, + "learning_rate": 1.6101768628587997e-05, + "loss": 0.3932, + "step": 2690 + }, + { + "epoch": 0.39143198869196477, + "grad_norm": 2.703125, + "learning_rate": 1.608727167294868e-05, + "loss": 0.4219, + "step": 2700 + }, + { + "epoch": 0.3928817367982313, + "grad_norm": 2.828125, + "learning_rate": 1.607277471730937e-05, + "loss": 0.4117, + "step": 2710 + }, + { + "epoch": 0.3943314849044978, + "grad_norm": 2.78125, + "learning_rate": 1.605827776167005e-05, + "loss": 0.3557, + "step": 2720 + }, + { + "epoch": 0.3957812330107644, + "grad_norm": 2.375, + "learning_rate": 1.6043780806030734e-05, + "loss": 0.3723, + "step": 2730 + }, + { + "epoch": 0.3972309811170309, + "grad_norm": 3.09375, + "learning_rate": 1.602928385039142e-05, + "loss": 0.3992, + "step": 2740 + }, + { + "epoch": 0.39868072922329745, + "grad_norm": 2.84375, + "learning_rate": 1.6014786894752103e-05, + "loss": 0.4101, + "step": 2750 + }, + { + "epoch": 0.40013047732956397, + "grad_norm": 2.40625, + "learning_rate": 1.6000289939112787e-05, + "loss": 0.4028, + "step": 2760 + }, + { + "epoch": 0.40158022543583055, + "grad_norm": 2.75, + "learning_rate": 1.5985792983473472e-05, + "loss": 0.4268, + "step": 2770 + }, + { + "epoch": 0.4030299735420971, + "grad_norm": 2.34375, + "learning_rate": 1.5971296027834156e-05, + "loss": 0.382, + "step": 2780 + }, + { + "epoch": 0.4044797216483636, + "grad_norm": 2.640625, + "learning_rate": 1.595679907219484e-05, + "loss": 0.4143, + "step": 2790 + }, + { + "epoch": 0.4059294697546301, + "grad_norm": 2.375, + "learning_rate": 1.5942302116555525e-05, + "loss": 0.3776, + "step": 2800 + }, + { + "epoch": 0.40737921786089665, + "grad_norm": 2.609375, + "learning_rate": 1.592780516091621e-05, + "loss": 0.3881, + "step": 2810 + }, + { + "epoch": 0.4088289659671632, + "grad_norm": 2.453125, + "learning_rate": 1.5913308205276894e-05, + "loss": 0.3945, + "step": 2820 + }, + { + "epoch": 0.41027871407342975, + "grad_norm": 2.21875, + "learning_rate": 1.589881124963758e-05, + "loss": 0.3863, + "step": 2830 + }, + { + "epoch": 0.4117284621796963, + "grad_norm": 2.40625, + "learning_rate": 1.5884314293998263e-05, + "loss": 0.3503, + "step": 2840 + }, + { + "epoch": 0.4131782102859628, + "grad_norm": 2.703125, + "learning_rate": 1.5869817338358947e-05, + "loss": 0.3564, + "step": 2850 + }, + { + "epoch": 0.4146279583922293, + "grad_norm": 2.484375, + "learning_rate": 1.585532038271963e-05, + "loss": 0.3654, + "step": 2860 + }, + { + "epoch": 0.4160777064984959, + "grad_norm": 2.8125, + "learning_rate": 1.5840823427080313e-05, + "loss": 0.41, + "step": 2870 + }, + { + "epoch": 0.4175274546047624, + "grad_norm": 2.65625, + "learning_rate": 1.5826326471440997e-05, + "loss": 0.3976, + "step": 2880 + }, + { + "epoch": 0.41897720271102895, + "grad_norm": 2.4375, + "learning_rate": 1.581182951580168e-05, + "loss": 0.3643, + "step": 2890 + }, + { + "epoch": 0.4204269508172955, + "grad_norm": 2.40625, + "learning_rate": 1.5797332560162366e-05, + "loss": 0.3715, + "step": 2900 + }, + { + "epoch": 0.42187669892356205, + "grad_norm": 2.59375, + "learning_rate": 1.5782835604523054e-05, + "loss": 0.3685, + "step": 2910 + }, + { + "epoch": 0.4233264470298286, + "grad_norm": 2.3125, + "learning_rate": 1.5768338648883735e-05, + "loss": 0.3913, + "step": 2920 + }, + { + "epoch": 0.4247761951360951, + "grad_norm": 2.484375, + "learning_rate": 1.575384169324442e-05, + "loss": 0.3755, + "step": 2930 + }, + { + "epoch": 0.4262259432423616, + "grad_norm": 2.21875, + "learning_rate": 1.5739344737605104e-05, + "loss": 0.3715, + "step": 2940 + }, + { + "epoch": 0.42767569134862815, + "grad_norm": 2.15625, + "learning_rate": 1.5724847781965788e-05, + "loss": 0.366, + "step": 2950 + }, + { + "epoch": 0.42912543945489473, + "grad_norm": 2.90625, + "learning_rate": 1.5710350826326472e-05, + "loss": 0.3655, + "step": 2960 + }, + { + "epoch": 0.43057518756116125, + "grad_norm": 2.59375, + "learning_rate": 1.5695853870687157e-05, + "loss": 0.3745, + "step": 2970 + }, + { + "epoch": 0.4320249356674278, + "grad_norm": 2.71875, + "learning_rate": 1.568135691504784e-05, + "loss": 0.3592, + "step": 2980 + }, + { + "epoch": 0.4334746837736943, + "grad_norm": 2.328125, + "learning_rate": 1.5666859959408526e-05, + "loss": 0.3933, + "step": 2990 + }, + { + "epoch": 0.4349244318799609, + "grad_norm": 2.46875, + "learning_rate": 1.565236300376921e-05, + "loss": 0.3456, + "step": 3000 + }, + { + "epoch": 0.4363741799862274, + "grad_norm": 2.53125, + "learning_rate": 1.5637866048129895e-05, + "loss": 0.343, + "step": 3010 + }, + { + "epoch": 0.43782392809249393, + "grad_norm": 2.296875, + "learning_rate": 1.562336909249058e-05, + "loss": 0.3819, + "step": 3020 + }, + { + "epoch": 0.43927367619876045, + "grad_norm": 2.375, + "learning_rate": 1.5608872136851263e-05, + "loss": 0.3599, + "step": 3030 + }, + { + "epoch": 0.440723424305027, + "grad_norm": 2.078125, + "learning_rate": 1.5594375181211948e-05, + "loss": 0.3384, + "step": 3040 + }, + { + "epoch": 0.44217317241129356, + "grad_norm": 2.375, + "learning_rate": 1.5579878225572632e-05, + "loss": 0.3701, + "step": 3050 + }, + { + "epoch": 0.4436229205175601, + "grad_norm": 2.828125, + "learning_rate": 1.5565381269933317e-05, + "loss": 0.3403, + "step": 3060 + }, + { + "epoch": 0.4450726686238266, + "grad_norm": 2.328125, + "learning_rate": 1.5550884314293998e-05, + "loss": 0.3721, + "step": 3070 + }, + { + "epoch": 0.44652241673009313, + "grad_norm": 2.640625, + "learning_rate": 1.5536387358654682e-05, + "loss": 0.3634, + "step": 3080 + }, + { + "epoch": 0.4479721648363597, + "grad_norm": 2.59375, + "learning_rate": 1.5521890403015366e-05, + "loss": 0.3548, + "step": 3090 + }, + { + "epoch": 0.44942191294262623, + "grad_norm": 2.328125, + "learning_rate": 1.550739344737605e-05, + "loss": 0.3467, + "step": 3100 + }, + { + "epoch": 0.45087166104889276, + "grad_norm": 2.484375, + "learning_rate": 1.549289649173674e-05, + "loss": 0.3681, + "step": 3110 + }, + { + "epoch": 0.4523214091551593, + "grad_norm": 2.203125, + "learning_rate": 1.5478399536097423e-05, + "loss": 0.3505, + "step": 3120 + }, + { + "epoch": 0.4537711572614258, + "grad_norm": 2.734375, + "learning_rate": 1.5463902580458104e-05, + "loss": 0.3434, + "step": 3130 + }, + { + "epoch": 0.4552209053676924, + "grad_norm": 2.1875, + "learning_rate": 1.544940562481879e-05, + "loss": 0.3699, + "step": 3140 + }, + { + "epoch": 0.4566706534739589, + "grad_norm": 2.46875, + "learning_rate": 1.5434908669179473e-05, + "loss": 0.3535, + "step": 3150 + }, + { + "epoch": 0.45812040158022543, + "grad_norm": 2.375, + "learning_rate": 1.5420411713540157e-05, + "loss": 0.347, + "step": 3160 + }, + { + "epoch": 0.45957014968649196, + "grad_norm": 2.4375, + "learning_rate": 1.5405914757900842e-05, + "loss": 0.3246, + "step": 3170 + }, + { + "epoch": 0.4610198977927585, + "grad_norm": 2.359375, + "learning_rate": 1.5391417802261526e-05, + "loss": 0.3855, + "step": 3180 + }, + { + "epoch": 0.46246964589902506, + "grad_norm": 2.703125, + "learning_rate": 1.537692084662221e-05, + "loss": 0.337, + "step": 3190 + }, + { + "epoch": 0.4639193940052916, + "grad_norm": 2.09375, + "learning_rate": 1.5362423890982895e-05, + "loss": 0.3662, + "step": 3200 + }, + { + "epoch": 0.4653691421115581, + "grad_norm": 2.421875, + "learning_rate": 1.534792693534358e-05, + "loss": 0.3239, + "step": 3210 + }, + { + "epoch": 0.46681889021782463, + "grad_norm": 2.328125, + "learning_rate": 1.5333429979704264e-05, + "loss": 0.3215, + "step": 3220 + }, + { + "epoch": 0.4682686383240912, + "grad_norm": 2.484375, + "learning_rate": 1.5318933024064948e-05, + "loss": 0.3532, + "step": 3230 + }, + { + "epoch": 0.46971838643035774, + "grad_norm": 2.1875, + "learning_rate": 1.5304436068425633e-05, + "loss": 0.3381, + "step": 3240 + }, + { + "epoch": 0.47116813453662426, + "grad_norm": 2.265625, + "learning_rate": 1.5289939112786317e-05, + "loss": 0.3519, + "step": 3250 + }, + { + "epoch": 0.4726178826428908, + "grad_norm": 2.234375, + "learning_rate": 1.5275442157147e-05, + "loss": 0.3229, + "step": 3260 + }, + { + "epoch": 0.4740676307491573, + "grad_norm": 2.203125, + "learning_rate": 1.5260945201507683e-05, + "loss": 0.3448, + "step": 3270 + }, + { + "epoch": 0.4755173788554239, + "grad_norm": 2.25, + "learning_rate": 1.5246448245868369e-05, + "loss": 0.3678, + "step": 3280 + }, + { + "epoch": 0.4769671269616904, + "grad_norm": 2.40625, + "learning_rate": 1.5231951290229053e-05, + "loss": 0.3244, + "step": 3290 + }, + { + "epoch": 0.47841687506795694, + "grad_norm": 2.546875, + "learning_rate": 1.5217454334589736e-05, + "loss": 0.3434, + "step": 3300 + }, + { + "epoch": 0.47986662317422346, + "grad_norm": 1.984375, + "learning_rate": 1.5202957378950422e-05, + "loss": 0.2983, + "step": 3310 + }, + { + "epoch": 0.48131637128049004, + "grad_norm": 2.71875, + "learning_rate": 1.5188460423311106e-05, + "loss": 0.3329, + "step": 3320 + }, + { + "epoch": 0.48276611938675656, + "grad_norm": 3.0625, + "learning_rate": 1.517396346767179e-05, + "loss": 0.3534, + "step": 3330 + }, + { + "epoch": 0.4842158674930231, + "grad_norm": 1.7890625, + "learning_rate": 1.5159466512032475e-05, + "loss": 0.3777, + "step": 3340 + }, + { + "epoch": 0.4856656155992896, + "grad_norm": 2.4375, + "learning_rate": 1.5144969556393158e-05, + "loss": 0.3539, + "step": 3350 + }, + { + "epoch": 0.48711536370555614, + "grad_norm": 2.46875, + "learning_rate": 1.5130472600753842e-05, + "loss": 0.2904, + "step": 3360 + }, + { + "epoch": 0.4885651118118227, + "grad_norm": 2.71875, + "learning_rate": 1.5115975645114527e-05, + "loss": 0.3442, + "step": 3370 + }, + { + "epoch": 0.49001485991808924, + "grad_norm": 2.25, + "learning_rate": 1.5101478689475211e-05, + "loss": 0.295, + "step": 3380 + }, + { + "epoch": 0.49146460802435576, + "grad_norm": 2.0625, + "learning_rate": 1.5086981733835896e-05, + "loss": 0.3025, + "step": 3390 + }, + { + "epoch": 0.4929143561306223, + "grad_norm": 2.421875, + "learning_rate": 1.5072484778196578e-05, + "loss": 0.3171, + "step": 3400 + }, + { + "epoch": 0.49436410423688887, + "grad_norm": 2.53125, + "learning_rate": 1.5057987822557264e-05, + "loss": 0.3019, + "step": 3410 + }, + { + "epoch": 0.4958138523431554, + "grad_norm": 1.9140625, + "learning_rate": 1.5043490866917949e-05, + "loss": 0.3502, + "step": 3420 + }, + { + "epoch": 0.4972636004494219, + "grad_norm": 2.109375, + "learning_rate": 1.5028993911278633e-05, + "loss": 0.3451, + "step": 3430 + }, + { + "epoch": 0.49871334855568844, + "grad_norm": 2.421875, + "learning_rate": 1.5014496955639318e-05, + "loss": 0.3151, + "step": 3440 + }, + { + "epoch": 0.500163096661955, + "grad_norm": 2.046875, + "learning_rate": 1.5000000000000002e-05, + "loss": 0.3123, + "step": 3450 + }, + { + "epoch": 0.5016128447682215, + "grad_norm": 2.484375, + "learning_rate": 1.4985503044360685e-05, + "loss": 0.3064, + "step": 3460 + }, + { + "epoch": 0.5030625928744881, + "grad_norm": 2.25, + "learning_rate": 1.497100608872137e-05, + "loss": 0.331, + "step": 3470 + }, + { + "epoch": 0.5045123409807546, + "grad_norm": 2.421875, + "learning_rate": 1.4956509133082054e-05, + "loss": 0.2976, + "step": 3480 + }, + { + "epoch": 0.5059620890870211, + "grad_norm": 2.125, + "learning_rate": 1.4942012177442738e-05, + "loss": 0.3207, + "step": 3490 + }, + { + "epoch": 0.5074118371932876, + "grad_norm": 2.0625, + "learning_rate": 1.4927515221803424e-05, + "loss": 0.2869, + "step": 3500 + }, + { + "epoch": 0.5088615852995542, + "grad_norm": 2.203125, + "learning_rate": 1.4913018266164107e-05, + "loss": 0.3411, + "step": 3510 + }, + { + "epoch": 0.5103113334058207, + "grad_norm": 1.78125, + "learning_rate": 1.4898521310524791e-05, + "loss": 0.264, + "step": 3520 + }, + { + "epoch": 0.5117610815120873, + "grad_norm": 2.453125, + "learning_rate": 1.4884024354885476e-05, + "loss": 0.3192, + "step": 3530 + }, + { + "epoch": 0.5132108296183538, + "grad_norm": 2.109375, + "learning_rate": 1.486952739924616e-05, + "loss": 0.2999, + "step": 3540 + }, + { + "epoch": 0.5146605777246204, + "grad_norm": 2.546875, + "learning_rate": 1.4855030443606845e-05, + "loss": 0.2922, + "step": 3550 + }, + { + "epoch": 0.5161103258308869, + "grad_norm": 2.015625, + "learning_rate": 1.4840533487967527e-05, + "loss": 0.329, + "step": 3560 + }, + { + "epoch": 0.5175600739371534, + "grad_norm": 2.25, + "learning_rate": 1.4826036532328212e-05, + "loss": 0.3185, + "step": 3570 + }, + { + "epoch": 0.5190098220434199, + "grad_norm": 2.296875, + "learning_rate": 1.4811539576688896e-05, + "loss": 0.2912, + "step": 3580 + }, + { + "epoch": 0.5204595701496865, + "grad_norm": 2.03125, + "learning_rate": 1.479704262104958e-05, + "loss": 0.2868, + "step": 3590 + }, + { + "epoch": 0.521909318255953, + "grad_norm": 2.5, + "learning_rate": 1.4782545665410267e-05, + "loss": 0.3219, + "step": 3600 + }, + { + "epoch": 0.5233590663622195, + "grad_norm": 2.171875, + "learning_rate": 1.476804870977095e-05, + "loss": 0.3063, + "step": 3610 + }, + { + "epoch": 0.5248088144684862, + "grad_norm": 1.90625, + "learning_rate": 1.4753551754131634e-05, + "loss": 0.3285, + "step": 3620 + }, + { + "epoch": 0.5262585625747527, + "grad_norm": 2.265625, + "learning_rate": 1.4739054798492318e-05, + "loss": 0.3328, + "step": 3630 + }, + { + "epoch": 0.5277083106810192, + "grad_norm": 2.140625, + "learning_rate": 1.4724557842853003e-05, + "loss": 0.3459, + "step": 3640 + }, + { + "epoch": 0.5291580587872857, + "grad_norm": 2.28125, + "learning_rate": 1.4710060887213687e-05, + "loss": 0.2812, + "step": 3650 + }, + { + "epoch": 0.5306078068935522, + "grad_norm": 1.84375, + "learning_rate": 1.469556393157437e-05, + "loss": 0.3076, + "step": 3660 + }, + { + "epoch": 0.5320575549998188, + "grad_norm": 1.984375, + "learning_rate": 1.4681066975935054e-05, + "loss": 0.2806, + "step": 3670 + }, + { + "epoch": 0.5335073031060853, + "grad_norm": 2.453125, + "learning_rate": 1.4666570020295739e-05, + "loss": 0.2855, + "step": 3680 + }, + { + "epoch": 0.5349570512123518, + "grad_norm": 1.9296875, + "learning_rate": 1.4652073064656423e-05, + "loss": 0.2886, + "step": 3690 + }, + { + "epoch": 0.5364067993186183, + "grad_norm": 1.984375, + "learning_rate": 1.4637576109017109e-05, + "loss": 0.2997, + "step": 3700 + }, + { + "epoch": 0.537856547424885, + "grad_norm": 2.1875, + "learning_rate": 1.4623079153377794e-05, + "loss": 0.3326, + "step": 3710 + }, + { + "epoch": 0.5393062955311515, + "grad_norm": 2.4375, + "learning_rate": 1.4608582197738476e-05, + "loss": 0.3038, + "step": 3720 + }, + { + "epoch": 0.540756043637418, + "grad_norm": 2.171875, + "learning_rate": 1.459408524209916e-05, + "loss": 0.277, + "step": 3730 + }, + { + "epoch": 0.5422057917436846, + "grad_norm": 2.015625, + "learning_rate": 1.4579588286459845e-05, + "loss": 0.2777, + "step": 3740 + }, + { + "epoch": 0.5436555398499511, + "grad_norm": 1.9921875, + "learning_rate": 1.456509133082053e-05, + "loss": 0.2858, + "step": 3750 + }, + { + "epoch": 0.5451052879562176, + "grad_norm": 1.9453125, + "learning_rate": 1.4550594375181212e-05, + "loss": 0.3235, + "step": 3760 + }, + { + "epoch": 0.5465550360624841, + "grad_norm": 2.390625, + "learning_rate": 1.4536097419541897e-05, + "loss": 0.3304, + "step": 3770 + }, + { + "epoch": 0.5480047841687506, + "grad_norm": 2.15625, + "learning_rate": 1.4521600463902581e-05, + "loss": 0.2676, + "step": 3780 + }, + { + "epoch": 0.5494545322750172, + "grad_norm": 2.265625, + "learning_rate": 1.4507103508263265e-05, + "loss": 0.2916, + "step": 3790 + }, + { + "epoch": 0.5509042803812838, + "grad_norm": 2.40625, + "learning_rate": 1.4492606552623952e-05, + "loss": 0.315, + "step": 3800 + }, + { + "epoch": 0.5523540284875503, + "grad_norm": 2.390625, + "learning_rate": 1.4478109596984636e-05, + "loss": 0.3041, + "step": 3810 + }, + { + "epoch": 0.5538037765938169, + "grad_norm": 2.03125, + "learning_rate": 1.4463612641345319e-05, + "loss": 0.2836, + "step": 3820 + }, + { + "epoch": 0.5552535247000834, + "grad_norm": 2.359375, + "learning_rate": 1.4449115685706003e-05, + "loss": 0.2971, + "step": 3830 + }, + { + "epoch": 0.5567032728063499, + "grad_norm": 2.265625, + "learning_rate": 1.4434618730066688e-05, + "loss": 0.2856, + "step": 3840 + }, + { + "epoch": 0.5581530209126164, + "grad_norm": 2.5625, + "learning_rate": 1.4420121774427372e-05, + "loss": 0.2874, + "step": 3850 + }, + { + "epoch": 0.559602769018883, + "grad_norm": 2.171875, + "learning_rate": 1.4405624818788055e-05, + "loss": 0.29, + "step": 3860 + }, + { + "epoch": 0.5610525171251495, + "grad_norm": 1.9609375, + "learning_rate": 1.4391127863148739e-05, + "loss": 0.2661, + "step": 3870 + }, + { + "epoch": 0.562502265231416, + "grad_norm": 2.109375, + "learning_rate": 1.4376630907509424e-05, + "loss": 0.3105, + "step": 3880 + }, + { + "epoch": 0.5639520133376825, + "grad_norm": 2.015625, + "learning_rate": 1.4362133951870108e-05, + "loss": 0.3044, + "step": 3890 + }, + { + "epoch": 0.5654017614439492, + "grad_norm": 1.7734375, + "learning_rate": 1.4347636996230794e-05, + "loss": 0.2948, + "step": 3900 + }, + { + "epoch": 0.5668515095502157, + "grad_norm": 2.03125, + "learning_rate": 1.4333140040591478e-05, + "loss": 0.302, + "step": 3910 + }, + { + "epoch": 0.5683012576564822, + "grad_norm": 2.21875, + "learning_rate": 1.4318643084952161e-05, + "loss": 0.2946, + "step": 3920 + }, + { + "epoch": 0.5697510057627487, + "grad_norm": 1.7890625, + "learning_rate": 1.4304146129312846e-05, + "loss": 0.3095, + "step": 3930 + }, + { + "epoch": 0.5712007538690153, + "grad_norm": 2.5625, + "learning_rate": 1.428964917367353e-05, + "loss": 0.3023, + "step": 3940 + }, + { + "epoch": 0.5726505019752818, + "grad_norm": 2.5, + "learning_rate": 1.4275152218034214e-05, + "loss": 0.2455, + "step": 3950 + }, + { + "epoch": 0.5741002500815483, + "grad_norm": 1.9296875, + "learning_rate": 1.4260655262394897e-05, + "loss": 0.2896, + "step": 3960 + }, + { + "epoch": 0.5755499981878148, + "grad_norm": 1.8671875, + "learning_rate": 1.4246158306755582e-05, + "loss": 0.2595, + "step": 3970 + }, + { + "epoch": 0.5769997462940814, + "grad_norm": 1.984375, + "learning_rate": 1.4231661351116266e-05, + "loss": 0.289, + "step": 3980 + }, + { + "epoch": 0.578449494400348, + "grad_norm": 2.109375, + "learning_rate": 1.421716439547695e-05, + "loss": 0.2666, + "step": 3990 + }, + { + "epoch": 0.5798992425066145, + "grad_norm": 2.125, + "learning_rate": 1.4202667439837637e-05, + "loss": 0.3198, + "step": 4000 + }, + { + "epoch": 0.581348990612881, + "grad_norm": 2.28125, + "learning_rate": 1.4188170484198321e-05, + "loss": 0.2666, + "step": 4010 + }, + { + "epoch": 0.5827987387191476, + "grad_norm": 2.09375, + "learning_rate": 1.4173673528559004e-05, + "loss": 0.2831, + "step": 4020 + }, + { + "epoch": 0.5842484868254141, + "grad_norm": 2.0, + "learning_rate": 1.4159176572919688e-05, + "loss": 0.2414, + "step": 4030 + }, + { + "epoch": 0.5856982349316806, + "grad_norm": 1.90625, + "learning_rate": 1.4144679617280372e-05, + "loss": 0.2973, + "step": 4040 + }, + { + "epoch": 0.5871479830379471, + "grad_norm": 2.203125, + "learning_rate": 1.4130182661641057e-05, + "loss": 0.2725, + "step": 4050 + }, + { + "epoch": 0.5885977311442137, + "grad_norm": 1.7265625, + "learning_rate": 1.4115685706001741e-05, + "loss": 0.2716, + "step": 4060 + }, + { + "epoch": 0.5900474792504802, + "grad_norm": 2.671875, + "learning_rate": 1.4101188750362424e-05, + "loss": 0.2977, + "step": 4070 + }, + { + "epoch": 0.5914972273567468, + "grad_norm": 2.171875, + "learning_rate": 1.4086691794723108e-05, + "loss": 0.251, + "step": 4080 + }, + { + "epoch": 0.5929469754630133, + "grad_norm": 2.4375, + "learning_rate": 1.4072194839083793e-05, + "loss": 0.3201, + "step": 4090 + }, + { + "epoch": 0.5943967235692799, + "grad_norm": 1.9609375, + "learning_rate": 1.4057697883444479e-05, + "loss": 0.2786, + "step": 4100 + }, + { + "epoch": 0.5958464716755464, + "grad_norm": 1.921875, + "learning_rate": 1.4043200927805163e-05, + "loss": 0.2939, + "step": 4110 + }, + { + "epoch": 0.5972962197818129, + "grad_norm": 1.8828125, + "learning_rate": 1.4028703972165846e-05, + "loss": 0.2243, + "step": 4120 + }, + { + "epoch": 0.5987459678880794, + "grad_norm": 1.859375, + "learning_rate": 1.401420701652653e-05, + "loss": 0.2625, + "step": 4130 + }, + { + "epoch": 0.600195715994346, + "grad_norm": 1.953125, + "learning_rate": 1.3999710060887215e-05, + "loss": 0.2566, + "step": 4140 + }, + { + "epoch": 0.6016454641006125, + "grad_norm": 2.1875, + "learning_rate": 1.39852131052479e-05, + "loss": 0.297, + "step": 4150 + }, + { + "epoch": 0.603095212206879, + "grad_norm": 1.8984375, + "learning_rate": 1.3970716149608584e-05, + "loss": 0.2866, + "step": 4160 + }, + { + "epoch": 0.6045449603131456, + "grad_norm": 2.109375, + "learning_rate": 1.3956219193969267e-05, + "loss": 0.2851, + "step": 4170 + }, + { + "epoch": 0.6059947084194122, + "grad_norm": 1.7421875, + "learning_rate": 1.3941722238329951e-05, + "loss": 0.2841, + "step": 4180 + }, + { + "epoch": 0.6074444565256787, + "grad_norm": 2.234375, + "learning_rate": 1.3927225282690635e-05, + "loss": 0.2626, + "step": 4190 + }, + { + "epoch": 0.6088942046319452, + "grad_norm": 1.875, + "learning_rate": 1.3912728327051321e-05, + "loss": 0.2509, + "step": 4200 + }, + { + "epoch": 0.6103439527382117, + "grad_norm": 2.171875, + "learning_rate": 1.3898231371412006e-05, + "loss": 0.2815, + "step": 4210 + }, + { + "epoch": 0.6117937008444783, + "grad_norm": 2.09375, + "learning_rate": 1.3883734415772689e-05, + "loss": 0.282, + "step": 4220 + }, + { + "epoch": 0.6132434489507448, + "grad_norm": 1.828125, + "learning_rate": 1.3869237460133373e-05, + "loss": 0.2869, + "step": 4230 + }, + { + "epoch": 0.6146931970570113, + "grad_norm": 1.765625, + "learning_rate": 1.3854740504494057e-05, + "loss": 0.2436, + "step": 4240 + }, + { + "epoch": 0.6161429451632778, + "grad_norm": 2.015625, + "learning_rate": 1.3840243548854742e-05, + "loss": 0.2385, + "step": 4250 + }, + { + "epoch": 0.6175926932695445, + "grad_norm": 2.171875, + "learning_rate": 1.3825746593215426e-05, + "loss": 0.2492, + "step": 4260 + }, + { + "epoch": 0.619042441375811, + "grad_norm": 1.9921875, + "learning_rate": 1.3811249637576109e-05, + "loss": 0.2652, + "step": 4270 + }, + { + "epoch": 0.6204921894820775, + "grad_norm": 2.5, + "learning_rate": 1.3796752681936793e-05, + "loss": 0.2758, + "step": 4280 + }, + { + "epoch": 0.621941937588344, + "grad_norm": 1.9609375, + "learning_rate": 1.3782255726297478e-05, + "loss": 0.2744, + "step": 4290 + }, + { + "epoch": 0.6233916856946106, + "grad_norm": 2.03125, + "learning_rate": 1.3767758770658164e-05, + "loss": 0.2917, + "step": 4300 + }, + { + "epoch": 0.6248414338008771, + "grad_norm": 2.578125, + "learning_rate": 1.3753261815018848e-05, + "loss": 0.3037, + "step": 4310 + }, + { + "epoch": 0.6262911819071436, + "grad_norm": 1.78125, + "learning_rate": 1.3738764859379533e-05, + "loss": 0.2106, + "step": 4320 + }, + { + "epoch": 0.6277409300134101, + "grad_norm": 2.109375, + "learning_rate": 1.3724267903740215e-05, + "loss": 0.2452, + "step": 4330 + }, + { + "epoch": 0.6291906781196767, + "grad_norm": 2.3125, + "learning_rate": 1.37097709481009e-05, + "loss": 0.3048, + "step": 4340 + }, + { + "epoch": 0.6306404262259433, + "grad_norm": 2.21875, + "learning_rate": 1.3695273992461584e-05, + "loss": 0.2839, + "step": 4350 + }, + { + "epoch": 0.6320901743322098, + "grad_norm": 2.03125, + "learning_rate": 1.3680777036822269e-05, + "loss": 0.2875, + "step": 4360 + }, + { + "epoch": 0.6335399224384763, + "grad_norm": 2.1875, + "learning_rate": 1.3666280081182951e-05, + "loss": 0.2686, + "step": 4370 + }, + { + "epoch": 0.6349896705447429, + "grad_norm": 2.296875, + "learning_rate": 1.3651783125543636e-05, + "loss": 0.279, + "step": 4380 + }, + { + "epoch": 0.6364394186510094, + "grad_norm": 2.15625, + "learning_rate": 1.363728616990432e-05, + "loss": 0.2709, + "step": 4390 + }, + { + "epoch": 0.6378891667572759, + "grad_norm": 1.859375, + "learning_rate": 1.3622789214265006e-05, + "loss": 0.2722, + "step": 4400 + }, + { + "epoch": 0.6393389148635424, + "grad_norm": 2.359375, + "learning_rate": 1.360829225862569e-05, + "loss": 0.2585, + "step": 4410 + }, + { + "epoch": 0.640788662969809, + "grad_norm": 2.3125, + "learning_rate": 1.3593795302986375e-05, + "loss": 0.3041, + "step": 4420 + }, + { + "epoch": 0.6422384110760755, + "grad_norm": 1.953125, + "learning_rate": 1.3579298347347058e-05, + "loss": 0.2763, + "step": 4430 + }, + { + "epoch": 0.6436881591823421, + "grad_norm": 1.8828125, + "learning_rate": 1.3564801391707742e-05, + "loss": 0.2839, + "step": 4440 + }, + { + "epoch": 0.6451379072886086, + "grad_norm": 2.34375, + "learning_rate": 1.3550304436068427e-05, + "loss": 0.2862, + "step": 4450 + }, + { + "epoch": 0.6465876553948752, + "grad_norm": 2.046875, + "learning_rate": 1.3535807480429111e-05, + "loss": 0.2549, + "step": 4460 + }, + { + "epoch": 0.6480374035011417, + "grad_norm": 2.21875, + "learning_rate": 1.3521310524789794e-05, + "loss": 0.3009, + "step": 4470 + }, + { + "epoch": 0.6494871516074082, + "grad_norm": 2.0, + "learning_rate": 1.3506813569150478e-05, + "loss": 0.2688, + "step": 4480 + }, + { + "epoch": 0.6509368997136747, + "grad_norm": 1.78125, + "learning_rate": 1.3492316613511163e-05, + "loss": 0.2365, + "step": 4490 + }, + { + "epoch": 0.6523866478199413, + "grad_norm": 1.8359375, + "learning_rate": 1.3477819657871849e-05, + "loss": 0.2597, + "step": 4500 + }, + { + "epoch": 0.6538363959262078, + "grad_norm": 1.8828125, + "learning_rate": 1.3463322702232533e-05, + "loss": 0.2769, + "step": 4510 + }, + { + "epoch": 0.6552861440324743, + "grad_norm": 1.953125, + "learning_rate": 1.3448825746593218e-05, + "loss": 0.308, + "step": 4520 + }, + { + "epoch": 0.6567358921387408, + "grad_norm": 2.375, + "learning_rate": 1.34343287909539e-05, + "loss": 0.2582, + "step": 4530 + }, + { + "epoch": 0.6581856402450075, + "grad_norm": 1.75, + "learning_rate": 1.3419831835314585e-05, + "loss": 0.2667, + "step": 4540 + }, + { + "epoch": 0.659635388351274, + "grad_norm": 1.859375, + "learning_rate": 1.340533487967527e-05, + "loss": 0.2508, + "step": 4550 + }, + { + "epoch": 0.6610851364575405, + "grad_norm": 1.9296875, + "learning_rate": 1.3390837924035954e-05, + "loss": 0.2845, + "step": 4560 + }, + { + "epoch": 0.662534884563807, + "grad_norm": 1.8671875, + "learning_rate": 1.3376340968396636e-05, + "loss": 0.2645, + "step": 4570 + }, + { + "epoch": 0.6639846326700736, + "grad_norm": 2.078125, + "learning_rate": 1.336184401275732e-05, + "loss": 0.242, + "step": 4580 + }, + { + "epoch": 0.6654343807763401, + "grad_norm": 2.25, + "learning_rate": 1.3347347057118005e-05, + "loss": 0.2298, + "step": 4590 + }, + { + "epoch": 0.6668841288826066, + "grad_norm": 2.125, + "learning_rate": 1.3332850101478691e-05, + "loss": 0.2543, + "step": 4600 + }, + { + "epoch": 0.6683338769888731, + "grad_norm": 1.8984375, + "learning_rate": 1.3318353145839376e-05, + "loss": 0.2471, + "step": 4610 + }, + { + "epoch": 0.6697836250951397, + "grad_norm": 2.0, + "learning_rate": 1.330385619020006e-05, + "loss": 0.2579, + "step": 4620 + }, + { + "epoch": 0.6712333732014063, + "grad_norm": 2.078125, + "learning_rate": 1.3289359234560743e-05, + "loss": 0.2651, + "step": 4630 + }, + { + "epoch": 0.6726831213076728, + "grad_norm": 1.96875, + "learning_rate": 1.3274862278921427e-05, + "loss": 0.2256, + "step": 4640 + }, + { + "epoch": 0.6741328694139394, + "grad_norm": 2.046875, + "learning_rate": 1.3260365323282112e-05, + "loss": 0.2677, + "step": 4650 + }, + { + "epoch": 0.6755826175202059, + "grad_norm": 1.71875, + "learning_rate": 1.3245868367642796e-05, + "loss": 0.2694, + "step": 4660 + }, + { + "epoch": 0.6770323656264724, + "grad_norm": 1.90625, + "learning_rate": 1.323137141200348e-05, + "loss": 0.2627, + "step": 4670 + }, + { + "epoch": 0.6784821137327389, + "grad_norm": 2.046875, + "learning_rate": 1.3216874456364163e-05, + "loss": 0.2708, + "step": 4680 + }, + { + "epoch": 0.6799318618390054, + "grad_norm": 2.046875, + "learning_rate": 1.3202377500724848e-05, + "loss": 0.2499, + "step": 4690 + }, + { + "epoch": 0.681381609945272, + "grad_norm": 1.6640625, + "learning_rate": 1.3187880545085534e-05, + "loss": 0.2368, + "step": 4700 + }, + { + "epoch": 0.6828313580515385, + "grad_norm": 1.8046875, + "learning_rate": 1.3173383589446218e-05, + "loss": 0.2566, + "step": 4710 + }, + { + "epoch": 0.6842811061578051, + "grad_norm": 1.9296875, + "learning_rate": 1.3158886633806903e-05, + "loss": 0.2607, + "step": 4720 + }, + { + "epoch": 0.6857308542640717, + "grad_norm": 1.9375, + "learning_rate": 1.3144389678167585e-05, + "loss": 0.2729, + "step": 4730 + }, + { + "epoch": 0.6871806023703382, + "grad_norm": 1.984375, + "learning_rate": 1.312989272252827e-05, + "loss": 0.2901, + "step": 4740 + }, + { + "epoch": 0.6886303504766047, + "grad_norm": 1.6015625, + "learning_rate": 1.3115395766888954e-05, + "loss": 0.2545, + "step": 4750 + }, + { + "epoch": 0.6900800985828712, + "grad_norm": 2.046875, + "learning_rate": 1.3100898811249639e-05, + "loss": 0.3032, + "step": 4760 + }, + { + "epoch": 0.6915298466891378, + "grad_norm": 1.609375, + "learning_rate": 1.3086401855610323e-05, + "loss": 0.248, + "step": 4770 + }, + { + "epoch": 0.6929795947954043, + "grad_norm": 1.953125, + "learning_rate": 1.3071904899971006e-05, + "loss": 0.2771, + "step": 4780 + }, + { + "epoch": 0.6944293429016708, + "grad_norm": 1.8125, + "learning_rate": 1.305740794433169e-05, + "loss": 0.2461, + "step": 4790 + }, + { + "epoch": 0.6958790910079373, + "grad_norm": 2.046875, + "learning_rate": 1.3042910988692376e-05, + "loss": 0.2745, + "step": 4800 + }, + { + "epoch": 0.697328839114204, + "grad_norm": 1.765625, + "learning_rate": 1.302841403305306e-05, + "loss": 0.2227, + "step": 4810 + }, + { + "epoch": 0.6987785872204705, + "grad_norm": 2.25, + "learning_rate": 1.3013917077413745e-05, + "loss": 0.2337, + "step": 4820 + }, + { + "epoch": 0.700228335326737, + "grad_norm": 2.265625, + "learning_rate": 1.2999420121774428e-05, + "loss": 0.2619, + "step": 4830 + }, + { + "epoch": 0.7016780834330035, + "grad_norm": 1.8359375, + "learning_rate": 1.2984923166135112e-05, + "loss": 0.2547, + "step": 4840 + }, + { + "epoch": 0.70312783153927, + "grad_norm": 2.140625, + "learning_rate": 1.2970426210495797e-05, + "loss": 0.2535, + "step": 4850 + }, + { + "epoch": 0.7045775796455366, + "grad_norm": 1.84375, + "learning_rate": 1.2955929254856481e-05, + "loss": 0.2413, + "step": 4860 + }, + { + "epoch": 0.7060273277518031, + "grad_norm": 1.875, + "learning_rate": 1.2941432299217166e-05, + "loss": 0.24, + "step": 4870 + }, + { + "epoch": 0.7074770758580696, + "grad_norm": 2.203125, + "learning_rate": 1.2926935343577848e-05, + "loss": 0.2506, + "step": 4880 + }, + { + "epoch": 0.7089268239643362, + "grad_norm": 2.203125, + "learning_rate": 1.2912438387938533e-05, + "loss": 0.2338, + "step": 4890 + }, + { + "epoch": 0.7103765720706028, + "grad_norm": 1.6875, + "learning_rate": 1.2897941432299219e-05, + "loss": 0.2733, + "step": 4900 + }, + { + "epoch": 0.7118263201768693, + "grad_norm": 2.046875, + "learning_rate": 1.2883444476659903e-05, + "loss": 0.2348, + "step": 4910 + }, + { + "epoch": 0.7132760682831358, + "grad_norm": 1.7265625, + "learning_rate": 1.2868947521020588e-05, + "loss": 0.2806, + "step": 4920 + }, + { + "epoch": 0.7147258163894024, + "grad_norm": 1.5703125, + "learning_rate": 1.2854450565381272e-05, + "loss": 0.2315, + "step": 4930 + }, + { + "epoch": 0.7161755644956689, + "grad_norm": 1.9921875, + "learning_rate": 1.2839953609741955e-05, + "loss": 0.2482, + "step": 4940 + }, + { + "epoch": 0.7176253126019354, + "grad_norm": 1.671875, + "learning_rate": 1.282545665410264e-05, + "loss": 0.2423, + "step": 4950 + }, + { + "epoch": 0.7190750607082019, + "grad_norm": 2.0625, + "learning_rate": 1.2810959698463324e-05, + "loss": 0.2637, + "step": 4960 + }, + { + "epoch": 0.7205248088144685, + "grad_norm": 1.90625, + "learning_rate": 1.2796462742824008e-05, + "loss": 0.1909, + "step": 4970 + }, + { + "epoch": 0.721974556920735, + "grad_norm": 2.0, + "learning_rate": 1.278196578718469e-05, + "loss": 0.2652, + "step": 4980 + }, + { + "epoch": 0.7234243050270016, + "grad_norm": 1.8828125, + "learning_rate": 1.2767468831545375e-05, + "loss": 0.2419, + "step": 4990 + }, + { + "epoch": 0.7248740531332681, + "grad_norm": 1.921875, + "learning_rate": 1.2752971875906061e-05, + "loss": 0.2425, + "step": 5000 + }, + { + "epoch": 0.7263238012395347, + "grad_norm": 1.78125, + "learning_rate": 1.2738474920266746e-05, + "loss": 0.2363, + "step": 5010 + }, + { + "epoch": 0.7277735493458012, + "grad_norm": 1.796875, + "learning_rate": 1.272397796462743e-05, + "loss": 0.2045, + "step": 5020 + }, + { + "epoch": 0.7292232974520677, + "grad_norm": 1.7734375, + "learning_rate": 1.2709481008988115e-05, + "loss": 0.236, + "step": 5030 + }, + { + "epoch": 0.7306730455583342, + "grad_norm": 1.7265625, + "learning_rate": 1.2694984053348797e-05, + "loss": 0.2325, + "step": 5040 + }, + { + "epoch": 0.7321227936646008, + "grad_norm": 1.4921875, + "learning_rate": 1.2680487097709482e-05, + "loss": 0.2425, + "step": 5050 + }, + { + "epoch": 0.7335725417708673, + "grad_norm": 1.640625, + "learning_rate": 1.2665990142070166e-05, + "loss": 0.2518, + "step": 5060 + }, + { + "epoch": 0.7350222898771338, + "grad_norm": 1.9375, + "learning_rate": 1.265149318643085e-05, + "loss": 0.2871, + "step": 5070 + }, + { + "epoch": 0.7364720379834003, + "grad_norm": 1.7265625, + "learning_rate": 1.2636996230791533e-05, + "loss": 0.2491, + "step": 5080 + }, + { + "epoch": 0.737921786089667, + "grad_norm": 2.03125, + "learning_rate": 1.2622499275152218e-05, + "loss": 0.2626, + "step": 5090 + }, + { + "epoch": 0.7393715341959335, + "grad_norm": 1.9453125, + "learning_rate": 1.2608002319512904e-05, + "loss": 0.28, + "step": 5100 + }, + { + "epoch": 0.7408212823022, + "grad_norm": 2.15625, + "learning_rate": 1.2593505363873588e-05, + "loss": 0.2514, + "step": 5110 + }, + { + "epoch": 0.7422710304084665, + "grad_norm": 2.015625, + "learning_rate": 1.2579008408234273e-05, + "loss": 0.2447, + "step": 5120 + }, + { + "epoch": 0.7437207785147331, + "grad_norm": 1.9375, + "learning_rate": 1.2564511452594957e-05, + "loss": 0.2955, + "step": 5130 + }, + { + "epoch": 0.7451705266209996, + "grad_norm": 1.875, + "learning_rate": 1.255001449695564e-05, + "loss": 0.2018, + "step": 5140 + }, + { + "epoch": 0.7466202747272661, + "grad_norm": 2.265625, + "learning_rate": 1.2535517541316324e-05, + "loss": 0.2646, + "step": 5150 + }, + { + "epoch": 0.7480700228335326, + "grad_norm": 1.8671875, + "learning_rate": 1.2521020585677009e-05, + "loss": 0.2455, + "step": 5160 + }, + { + "epoch": 0.7495197709397992, + "grad_norm": 1.8515625, + "learning_rate": 1.2506523630037693e-05, + "loss": 0.2129, + "step": 5170 + }, + { + "epoch": 0.7509695190460658, + "grad_norm": 1.8828125, + "learning_rate": 1.2492026674398376e-05, + "loss": 0.2461, + "step": 5180 + }, + { + "epoch": 0.7524192671523323, + "grad_norm": 2.03125, + "learning_rate": 1.2477529718759063e-05, + "loss": 0.2581, + "step": 5190 + }, + { + "epoch": 0.7538690152585988, + "grad_norm": 1.7734375, + "learning_rate": 1.2463032763119746e-05, + "loss": 0.232, + "step": 5200 + }, + { + "epoch": 0.7553187633648654, + "grad_norm": 2.4375, + "learning_rate": 1.244853580748043e-05, + "loss": 0.2385, + "step": 5210 + }, + { + "epoch": 0.7567685114711319, + "grad_norm": 1.515625, + "learning_rate": 1.2434038851841115e-05, + "loss": 0.2348, + "step": 5220 + }, + { + "epoch": 0.7582182595773984, + "grad_norm": 1.4765625, + "learning_rate": 1.24195418962018e-05, + "loss": 0.2408, + "step": 5230 + }, + { + "epoch": 0.7596680076836649, + "grad_norm": 2.0, + "learning_rate": 1.2405044940562482e-05, + "loss": 0.2063, + "step": 5240 + }, + { + "epoch": 0.7611177557899315, + "grad_norm": 2.015625, + "learning_rate": 1.2390547984923167e-05, + "loss": 0.2283, + "step": 5250 + }, + { + "epoch": 0.762567503896198, + "grad_norm": 2.09375, + "learning_rate": 1.2376051029283851e-05, + "loss": 0.2586, + "step": 5260 + }, + { + "epoch": 0.7640172520024646, + "grad_norm": 1.6875, + "learning_rate": 1.2361554073644535e-05, + "loss": 0.2276, + "step": 5270 + }, + { + "epoch": 0.7654670001087311, + "grad_norm": 1.9296875, + "learning_rate": 1.234705711800522e-05, + "loss": 0.2136, + "step": 5280 + }, + { + "epoch": 0.7669167482149977, + "grad_norm": 1.96875, + "learning_rate": 1.2332560162365906e-05, + "loss": 0.2388, + "step": 5290 + }, + { + "epoch": 0.7683664963212642, + "grad_norm": 2.015625, + "learning_rate": 1.2318063206726589e-05, + "loss": 0.2481, + "step": 5300 + }, + { + "epoch": 0.7698162444275307, + "grad_norm": 1.8046875, + "learning_rate": 1.2303566251087273e-05, + "loss": 0.2364, + "step": 5310 + }, + { + "epoch": 0.7712659925337972, + "grad_norm": 1.78125, + "learning_rate": 1.2289069295447958e-05, + "loss": 0.2209, + "step": 5320 + }, + { + "epoch": 0.7727157406400638, + "grad_norm": 1.96875, + "learning_rate": 1.2274572339808642e-05, + "loss": 0.2342, + "step": 5330 + }, + { + "epoch": 0.7741654887463303, + "grad_norm": 1.6015625, + "learning_rate": 1.2260075384169325e-05, + "loss": 0.2444, + "step": 5340 + }, + { + "epoch": 0.7756152368525968, + "grad_norm": 1.75, + "learning_rate": 1.2245578428530009e-05, + "loss": 0.2056, + "step": 5350 + }, + { + "epoch": 0.7770649849588634, + "grad_norm": 1.9375, + "learning_rate": 1.2231081472890693e-05, + "loss": 0.2596, + "step": 5360 + }, + { + "epoch": 0.77851473306513, + "grad_norm": 1.390625, + "learning_rate": 1.2216584517251378e-05, + "loss": 0.2151, + "step": 5370 + }, + { + "epoch": 0.7799644811713965, + "grad_norm": 1.7109375, + "learning_rate": 1.2202087561612062e-05, + "loss": 0.2188, + "step": 5380 + }, + { + "epoch": 0.781414229277663, + "grad_norm": 1.8984375, + "learning_rate": 1.2187590605972748e-05, + "loss": 0.2514, + "step": 5390 + }, + { + "epoch": 0.7828639773839295, + "grad_norm": 1.7734375, + "learning_rate": 1.2173093650333431e-05, + "loss": 0.2026, + "step": 5400 + }, + { + "epoch": 0.7843137254901961, + "grad_norm": 1.7734375, + "learning_rate": 1.2158596694694116e-05, + "loss": 0.2261, + "step": 5410 + }, + { + "epoch": 0.7857634735964626, + "grad_norm": 1.796875, + "learning_rate": 1.21440997390548e-05, + "loss": 0.219, + "step": 5420 + }, + { + "epoch": 0.7872132217027291, + "grad_norm": 1.96875, + "learning_rate": 1.2129602783415484e-05, + "loss": 0.2468, + "step": 5430 + }, + { + "epoch": 0.7886629698089956, + "grad_norm": 1.875, + "learning_rate": 1.2115105827776167e-05, + "loss": 0.2486, + "step": 5440 + }, + { + "epoch": 0.7901127179152623, + "grad_norm": 1.8984375, + "learning_rate": 1.2100608872136852e-05, + "loss": 0.2162, + "step": 5450 + }, + { + "epoch": 0.7915624660215288, + "grad_norm": 2.328125, + "learning_rate": 1.2086111916497536e-05, + "loss": 0.2382, + "step": 5460 + }, + { + "epoch": 0.7930122141277953, + "grad_norm": 1.71875, + "learning_rate": 1.207161496085822e-05, + "loss": 0.2424, + "step": 5470 + }, + { + "epoch": 0.7944619622340618, + "grad_norm": 2.328125, + "learning_rate": 1.2057118005218905e-05, + "loss": 0.2417, + "step": 5480 + }, + { + "epoch": 0.7959117103403284, + "grad_norm": 2.515625, + "learning_rate": 1.2042621049579591e-05, + "loss": 0.2332, + "step": 5490 + }, + { + "epoch": 0.7973614584465949, + "grad_norm": 1.9375, + "learning_rate": 1.2028124093940274e-05, + "loss": 0.2265, + "step": 5500 + }, + { + "epoch": 0.7988112065528614, + "grad_norm": 2.015625, + "learning_rate": 1.2013627138300958e-05, + "loss": 0.224, + "step": 5510 + }, + { + "epoch": 0.8002609546591279, + "grad_norm": 1.9375, + "learning_rate": 1.1999130182661642e-05, + "loss": 0.2264, + "step": 5520 + }, + { + "epoch": 0.8017107027653945, + "grad_norm": 1.90625, + "learning_rate": 1.1984633227022327e-05, + "loss": 0.2563, + "step": 5530 + }, + { + "epoch": 0.8031604508716611, + "grad_norm": 1.9453125, + "learning_rate": 1.1970136271383011e-05, + "loss": 0.2239, + "step": 5540 + }, + { + "epoch": 0.8046101989779276, + "grad_norm": 1.625, + "learning_rate": 1.1955639315743694e-05, + "loss": 0.2084, + "step": 5550 + }, + { + "epoch": 0.8060599470841942, + "grad_norm": 1.6640625, + "learning_rate": 1.1941142360104378e-05, + "loss": 0.2408, + "step": 5560 + }, + { + "epoch": 0.8075096951904607, + "grad_norm": 1.671875, + "learning_rate": 1.1926645404465063e-05, + "loss": 0.2034, + "step": 5570 + }, + { + "epoch": 0.8089594432967272, + "grad_norm": 2.140625, + "learning_rate": 1.1912148448825747e-05, + "loss": 0.2504, + "step": 5580 + }, + { + "epoch": 0.8104091914029937, + "grad_norm": 1.765625, + "learning_rate": 1.1897651493186433e-05, + "loss": 0.2253, + "step": 5590 + }, + { + "epoch": 0.8118589395092602, + "grad_norm": 2.703125, + "learning_rate": 1.1883154537547116e-05, + "loss": 0.2498, + "step": 5600 + }, + { + "epoch": 0.8133086876155268, + "grad_norm": 1.6171875, + "learning_rate": 1.18686575819078e-05, + "loss": 0.2267, + "step": 5610 + }, + { + "epoch": 0.8147584357217933, + "grad_norm": 1.5546875, + "learning_rate": 1.1854160626268485e-05, + "loss": 0.2274, + "step": 5620 + }, + { + "epoch": 0.8162081838280599, + "grad_norm": 1.96875, + "learning_rate": 1.183966367062917e-05, + "loss": 0.2231, + "step": 5630 + }, + { + "epoch": 0.8176579319343265, + "grad_norm": 1.921875, + "learning_rate": 1.1825166714989854e-05, + "loss": 0.2096, + "step": 5640 + }, + { + "epoch": 0.819107680040593, + "grad_norm": 1.7265625, + "learning_rate": 1.1810669759350536e-05, + "loss": 0.2704, + "step": 5650 + }, + { + "epoch": 0.8205574281468595, + "grad_norm": 1.6875, + "learning_rate": 1.1796172803711221e-05, + "loss": 0.2066, + "step": 5660 + }, + { + "epoch": 0.822007176253126, + "grad_norm": 1.84375, + "learning_rate": 1.1781675848071905e-05, + "loss": 0.2283, + "step": 5670 + }, + { + "epoch": 0.8234569243593926, + "grad_norm": 1.6875, + "learning_rate": 1.176717889243259e-05, + "loss": 0.2513, + "step": 5680 + }, + { + "epoch": 0.8249066724656591, + "grad_norm": 1.7890625, + "learning_rate": 1.1752681936793276e-05, + "loss": 0.2303, + "step": 5690 + }, + { + "epoch": 0.8263564205719256, + "grad_norm": 1.90625, + "learning_rate": 1.173818498115396e-05, + "loss": 0.2601, + "step": 5700 + }, + { + "epoch": 0.8278061686781921, + "grad_norm": 1.8515625, + "learning_rate": 1.1723688025514643e-05, + "loss": 0.2538, + "step": 5710 + }, + { + "epoch": 0.8292559167844586, + "grad_norm": 1.7734375, + "learning_rate": 1.1709191069875327e-05, + "loss": 0.2514, + "step": 5720 + }, + { + "epoch": 0.8307056648907253, + "grad_norm": 2.046875, + "learning_rate": 1.1694694114236012e-05, + "loss": 0.2452, + "step": 5730 + }, + { + "epoch": 0.8321554129969918, + "grad_norm": 1.7265625, + "learning_rate": 1.1680197158596696e-05, + "loss": 0.2453, + "step": 5740 + }, + { + "epoch": 0.8336051611032583, + "grad_norm": 2.078125, + "learning_rate": 1.1665700202957379e-05, + "loss": 0.2355, + "step": 5750 + }, + { + "epoch": 0.8350549092095249, + "grad_norm": 1.703125, + "learning_rate": 1.1651203247318063e-05, + "loss": 0.2142, + "step": 5760 + }, + { + "epoch": 0.8365046573157914, + "grad_norm": 2.078125, + "learning_rate": 1.1636706291678748e-05, + "loss": 0.2166, + "step": 5770 + }, + { + "epoch": 0.8379544054220579, + "grad_norm": 1.6796875, + "learning_rate": 1.1622209336039432e-05, + "loss": 0.2086, + "step": 5780 + }, + { + "epoch": 0.8394041535283244, + "grad_norm": 1.7578125, + "learning_rate": 1.1607712380400118e-05, + "loss": 0.2362, + "step": 5790 + }, + { + "epoch": 0.840853901634591, + "grad_norm": 1.9609375, + "learning_rate": 1.1593215424760803e-05, + "loss": 0.2538, + "step": 5800 + }, + { + "epoch": 0.8423036497408575, + "grad_norm": 1.8515625, + "learning_rate": 1.1578718469121485e-05, + "loss": 0.2435, + "step": 5810 + }, + { + "epoch": 0.8437533978471241, + "grad_norm": 2.0625, + "learning_rate": 1.156422151348217e-05, + "loss": 0.2312, + "step": 5820 + }, + { + "epoch": 0.8452031459533906, + "grad_norm": 2.125, + "learning_rate": 1.1549724557842854e-05, + "loss": 0.2263, + "step": 5830 + }, + { + "epoch": 0.8466528940596572, + "grad_norm": 1.9453125, + "learning_rate": 1.1535227602203539e-05, + "loss": 0.2159, + "step": 5840 + }, + { + "epoch": 0.8481026421659237, + "grad_norm": 1.578125, + "learning_rate": 1.1520730646564221e-05, + "loss": 0.2498, + "step": 5850 + }, + { + "epoch": 0.8495523902721902, + "grad_norm": 1.75, + "learning_rate": 1.1506233690924906e-05, + "loss": 0.2543, + "step": 5860 + }, + { + "epoch": 0.8510021383784567, + "grad_norm": 1.9140625, + "learning_rate": 1.149173673528559e-05, + "loss": 0.2424, + "step": 5870 + }, + { + "epoch": 0.8524518864847233, + "grad_norm": 2.125, + "learning_rate": 1.1477239779646275e-05, + "loss": 0.2203, + "step": 5880 + }, + { + "epoch": 0.8539016345909898, + "grad_norm": 1.640625, + "learning_rate": 1.146274282400696e-05, + "loss": 0.2512, + "step": 5890 + }, + { + "epoch": 0.8553513826972563, + "grad_norm": 1.734375, + "learning_rate": 1.1448245868367645e-05, + "loss": 0.2151, + "step": 5900 + }, + { + "epoch": 0.8568011308035229, + "grad_norm": 2.171875, + "learning_rate": 1.1433748912728328e-05, + "loss": 0.25, + "step": 5910 + }, + { + "epoch": 0.8582508789097895, + "grad_norm": 1.875, + "learning_rate": 1.1419251957089012e-05, + "loss": 0.2649, + "step": 5920 + }, + { + "epoch": 0.859700627016056, + "grad_norm": 1.6875, + "learning_rate": 1.1404755001449697e-05, + "loss": 0.223, + "step": 5930 + }, + { + "epoch": 0.8611503751223225, + "grad_norm": 1.734375, + "learning_rate": 1.1390258045810381e-05, + "loss": 0.2157, + "step": 5940 + }, + { + "epoch": 0.862600123228589, + "grad_norm": 1.4609375, + "learning_rate": 1.1375761090171064e-05, + "loss": 0.1802, + "step": 5950 + }, + { + "epoch": 0.8640498713348556, + "grad_norm": 1.6484375, + "learning_rate": 1.1361264134531748e-05, + "loss": 0.2185, + "step": 5960 + }, + { + "epoch": 0.8654996194411221, + "grad_norm": 1.640625, + "learning_rate": 1.1346767178892433e-05, + "loss": 0.2055, + "step": 5970 + }, + { + "epoch": 0.8669493675473886, + "grad_norm": 2.1875, + "learning_rate": 1.1332270223253117e-05, + "loss": 0.2157, + "step": 5980 + }, + { + "epoch": 0.8683991156536551, + "grad_norm": 1.9453125, + "learning_rate": 1.1317773267613803e-05, + "loss": 0.2298, + "step": 5990 + }, + { + "epoch": 0.8698488637599218, + "grad_norm": 1.71875, + "learning_rate": 1.1303276311974488e-05, + "loss": 0.2106, + "step": 6000 + }, + { + "epoch": 0.8712986118661883, + "grad_norm": 1.671875, + "learning_rate": 1.128877935633517e-05, + "loss": 0.2093, + "step": 6010 + }, + { + "epoch": 0.8727483599724548, + "grad_norm": 1.90625, + "learning_rate": 1.1274282400695855e-05, + "loss": 0.2418, + "step": 6020 + }, + { + "epoch": 0.8741981080787213, + "grad_norm": 2.125, + "learning_rate": 1.125978544505654e-05, + "loss": 0.2467, + "step": 6030 + }, + { + "epoch": 0.8756478561849879, + "grad_norm": 2.03125, + "learning_rate": 1.1245288489417224e-05, + "loss": 0.2458, + "step": 6040 + }, + { + "epoch": 0.8770976042912544, + "grad_norm": 1.8046875, + "learning_rate": 1.1230791533777906e-05, + "loss": 0.2425, + "step": 6050 + }, + { + "epoch": 0.8785473523975209, + "grad_norm": 1.5234375, + "learning_rate": 1.121629457813859e-05, + "loss": 0.2363, + "step": 6060 + }, + { + "epoch": 0.8799971005037874, + "grad_norm": 1.796875, + "learning_rate": 1.1201797622499275e-05, + "loss": 0.2397, + "step": 6070 + }, + { + "epoch": 0.881446848610054, + "grad_norm": 1.5390625, + "learning_rate": 1.118730066685996e-05, + "loss": 0.2139, + "step": 6080 + }, + { + "epoch": 0.8828965967163206, + "grad_norm": 1.6796875, + "learning_rate": 1.1172803711220646e-05, + "loss": 0.229, + "step": 6090 + }, + { + "epoch": 0.8843463448225871, + "grad_norm": 1.7109375, + "learning_rate": 1.115830675558133e-05, + "loss": 0.2004, + "step": 6100 + }, + { + "epoch": 0.8857960929288536, + "grad_norm": 1.890625, + "learning_rate": 1.1143809799942013e-05, + "loss": 0.2233, + "step": 6110 + }, + { + "epoch": 0.8872458410351202, + "grad_norm": 1.8359375, + "learning_rate": 1.1129312844302697e-05, + "loss": 0.2162, + "step": 6120 + }, + { + "epoch": 0.8886955891413867, + "grad_norm": 1.59375, + "learning_rate": 1.1114815888663382e-05, + "loss": 0.2619, + "step": 6130 + }, + { + "epoch": 0.8901453372476532, + "grad_norm": 1.84375, + "learning_rate": 1.1100318933024066e-05, + "loss": 0.2318, + "step": 6140 + }, + { + "epoch": 0.8915950853539197, + "grad_norm": 2.15625, + "learning_rate": 1.108582197738475e-05, + "loss": 0.2036, + "step": 6150 + }, + { + "epoch": 0.8930448334601863, + "grad_norm": 1.4765625, + "learning_rate": 1.1071325021745433e-05, + "loss": 0.2361, + "step": 6160 + }, + { + "epoch": 0.8944945815664528, + "grad_norm": 1.859375, + "learning_rate": 1.1056828066106118e-05, + "loss": 0.2239, + "step": 6170 + }, + { + "epoch": 0.8959443296727194, + "grad_norm": 1.75, + "learning_rate": 1.1042331110466802e-05, + "loss": 0.2337, + "step": 6180 + }, + { + "epoch": 0.8973940777789859, + "grad_norm": 2.21875, + "learning_rate": 1.1027834154827488e-05, + "loss": 0.23, + "step": 6190 + }, + { + "epoch": 0.8988438258852525, + "grad_norm": 1.734375, + "learning_rate": 1.1013337199188173e-05, + "loss": 0.2014, + "step": 6200 + }, + { + "epoch": 0.900293573991519, + "grad_norm": 1.5859375, + "learning_rate": 1.0998840243548855e-05, + "loss": 0.2323, + "step": 6210 + }, + { + "epoch": 0.9017433220977855, + "grad_norm": 1.296875, + "learning_rate": 1.098434328790954e-05, + "loss": 0.1748, + "step": 6220 + }, + { + "epoch": 0.903193070204052, + "grad_norm": 2.09375, + "learning_rate": 1.0969846332270224e-05, + "loss": 0.2295, + "step": 6230 + }, + { + "epoch": 0.9046428183103186, + "grad_norm": 1.40625, + "learning_rate": 1.0955349376630909e-05, + "loss": 0.2217, + "step": 6240 + }, + { + "epoch": 0.9060925664165851, + "grad_norm": 1.890625, + "learning_rate": 1.0940852420991593e-05, + "loss": 0.197, + "step": 6250 + }, + { + "epoch": 0.9075423145228516, + "grad_norm": 2.0, + "learning_rate": 1.0926355465352276e-05, + "loss": 0.2455, + "step": 6260 + }, + { + "epoch": 0.9089920626291181, + "grad_norm": 1.578125, + "learning_rate": 1.091185850971296e-05, + "loss": 0.1941, + "step": 6270 + }, + { + "epoch": 0.9104418107353848, + "grad_norm": 1.9296875, + "learning_rate": 1.0897361554073645e-05, + "loss": 0.2331, + "step": 6280 + }, + { + "epoch": 0.9118915588416513, + "grad_norm": 1.59375, + "learning_rate": 1.088286459843433e-05, + "loss": 0.2084, + "step": 6290 + }, + { + "epoch": 0.9133413069479178, + "grad_norm": 1.7421875, + "learning_rate": 1.0868367642795015e-05, + "loss": 0.2208, + "step": 6300 + }, + { + "epoch": 0.9147910550541843, + "grad_norm": 2.265625, + "learning_rate": 1.08538706871557e-05, + "loss": 0.2248, + "step": 6310 + }, + { + "epoch": 0.9162408031604509, + "grad_norm": 1.7421875, + "learning_rate": 1.0839373731516382e-05, + "loss": 0.2357, + "step": 6320 + }, + { + "epoch": 0.9176905512667174, + "grad_norm": 1.375, + "learning_rate": 1.0824876775877067e-05, + "loss": 0.2225, + "step": 6330 + }, + { + "epoch": 0.9191402993729839, + "grad_norm": 1.7578125, + "learning_rate": 1.0810379820237751e-05, + "loss": 0.1926, + "step": 6340 + }, + { + "epoch": 0.9205900474792504, + "grad_norm": 2.0625, + "learning_rate": 1.0795882864598435e-05, + "loss": 0.1965, + "step": 6350 + }, + { + "epoch": 0.922039795585517, + "grad_norm": 2.203125, + "learning_rate": 1.0781385908959118e-05, + "loss": 0.2076, + "step": 6360 + }, + { + "epoch": 0.9234895436917836, + "grad_norm": 1.65625, + "learning_rate": 1.0766888953319803e-05, + "loss": 0.2054, + "step": 6370 + }, + { + "epoch": 0.9249392917980501, + "grad_norm": 1.515625, + "learning_rate": 1.0752391997680487e-05, + "loss": 0.2021, + "step": 6380 + }, + { + "epoch": 0.9263890399043166, + "grad_norm": 1.484375, + "learning_rate": 1.0737895042041173e-05, + "loss": 0.1864, + "step": 6390 + }, + { + "epoch": 0.9278387880105832, + "grad_norm": 2.28125, + "learning_rate": 1.0723398086401858e-05, + "loss": 0.1947, + "step": 6400 + }, + { + "epoch": 0.9292885361168497, + "grad_norm": 1.7109375, + "learning_rate": 1.0708901130762542e-05, + "loss": 0.2233, + "step": 6410 + }, + { + "epoch": 0.9307382842231162, + "grad_norm": 1.671875, + "learning_rate": 1.0694404175123225e-05, + "loss": 0.2042, + "step": 6420 + }, + { + "epoch": 0.9321880323293827, + "grad_norm": 1.9609375, + "learning_rate": 1.0679907219483909e-05, + "loss": 0.1938, + "step": 6430 + }, + { + "epoch": 0.9336377804356493, + "grad_norm": 1.8984375, + "learning_rate": 1.0665410263844594e-05, + "loss": 0.2292, + "step": 6440 + }, + { + "epoch": 0.9350875285419158, + "grad_norm": 1.859375, + "learning_rate": 1.0650913308205278e-05, + "loss": 0.188, + "step": 6450 + }, + { + "epoch": 0.9365372766481824, + "grad_norm": 1.859375, + "learning_rate": 1.063641635256596e-05, + "loss": 0.2111, + "step": 6460 + }, + { + "epoch": 0.937987024754449, + "grad_norm": 1.4296875, + "learning_rate": 1.0621919396926645e-05, + "loss": 0.1935, + "step": 6470 + }, + { + "epoch": 0.9394367728607155, + "grad_norm": 1.65625, + "learning_rate": 1.060742244128733e-05, + "loss": 0.2337, + "step": 6480 + }, + { + "epoch": 0.940886520966982, + "grad_norm": 1.6171875, + "learning_rate": 1.0592925485648016e-05, + "loss": 0.217, + "step": 6490 + }, + { + "epoch": 0.9423362690732485, + "grad_norm": 1.640625, + "learning_rate": 1.05784285300087e-05, + "loss": 0.1999, + "step": 6500 + }, + { + "epoch": 0.943786017179515, + "grad_norm": 1.3984375, + "learning_rate": 1.0563931574369384e-05, + "loss": 0.2368, + "step": 6510 + }, + { + "epoch": 0.9452357652857816, + "grad_norm": 1.8515625, + "learning_rate": 1.0549434618730067e-05, + "loss": 0.2015, + "step": 6520 + }, + { + "epoch": 0.9466855133920481, + "grad_norm": 1.28125, + "learning_rate": 1.0534937663090752e-05, + "loss": 0.1905, + "step": 6530 + }, + { + "epoch": 0.9481352614983146, + "grad_norm": 1.6640625, + "learning_rate": 1.0520440707451436e-05, + "loss": 0.211, + "step": 6540 + }, + { + "epoch": 0.9495850096045813, + "grad_norm": 1.7421875, + "learning_rate": 1.050594375181212e-05, + "loss": 0.218, + "step": 6550 + }, + { + "epoch": 0.9510347577108478, + "grad_norm": 1.84375, + "learning_rate": 1.0491446796172803e-05, + "loss": 0.2278, + "step": 6560 + }, + { + "epoch": 0.9524845058171143, + "grad_norm": 1.5625, + "learning_rate": 1.0476949840533488e-05, + "loss": 0.2265, + "step": 6570 + }, + { + "epoch": 0.9539342539233808, + "grad_norm": 1.65625, + "learning_rate": 1.0462452884894172e-05, + "loss": 0.2122, + "step": 6580 + }, + { + "epoch": 0.9553840020296474, + "grad_norm": 2.203125, + "learning_rate": 1.0447955929254858e-05, + "loss": 0.1977, + "step": 6590 + }, + { + "epoch": 0.9568337501359139, + "grad_norm": 1.703125, + "learning_rate": 1.0433458973615543e-05, + "loss": 0.2086, + "step": 6600 + }, + { + "epoch": 0.9582834982421804, + "grad_norm": 1.46875, + "learning_rate": 1.0418962017976227e-05, + "loss": 0.2153, + "step": 6610 + }, + { + "epoch": 0.9597332463484469, + "grad_norm": 1.703125, + "learning_rate": 1.040446506233691e-05, + "loss": 0.1882, + "step": 6620 + }, + { + "epoch": 0.9611829944547134, + "grad_norm": 1.921875, + "learning_rate": 1.0389968106697594e-05, + "loss": 0.2172, + "step": 6630 + }, + { + "epoch": 0.9626327425609801, + "grad_norm": 1.453125, + "learning_rate": 1.0375471151058278e-05, + "loss": 0.1949, + "step": 6640 + }, + { + "epoch": 0.9640824906672466, + "grad_norm": 1.40625, + "learning_rate": 1.0360974195418963e-05, + "loss": 0.2007, + "step": 6650 + }, + { + "epoch": 0.9655322387735131, + "grad_norm": 2.203125, + "learning_rate": 1.0346477239779646e-05, + "loss": 0.2231, + "step": 6660 + }, + { + "epoch": 0.9669819868797797, + "grad_norm": 1.234375, + "learning_rate": 1.033198028414033e-05, + "loss": 0.1793, + "step": 6670 + }, + { + "epoch": 0.9684317349860462, + "grad_norm": 1.703125, + "learning_rate": 1.0317483328501014e-05, + "loss": 0.2123, + "step": 6680 + }, + { + "epoch": 0.9698814830923127, + "grad_norm": 1.734375, + "learning_rate": 1.03029863728617e-05, + "loss": 0.2037, + "step": 6690 + }, + { + "epoch": 0.9713312311985792, + "grad_norm": 1.796875, + "learning_rate": 1.0288489417222385e-05, + "loss": 0.1833, + "step": 6700 + }, + { + "epoch": 0.9727809793048457, + "grad_norm": 2.21875, + "learning_rate": 1.027399246158307e-05, + "loss": 0.2333, + "step": 6710 + }, + { + "epoch": 0.9742307274111123, + "grad_norm": 1.6875, + "learning_rate": 1.0259495505943752e-05, + "loss": 0.2051, + "step": 6720 + }, + { + "epoch": 0.9756804755173789, + "grad_norm": 1.671875, + "learning_rate": 1.0244998550304437e-05, + "loss": 0.2282, + "step": 6730 + }, + { + "epoch": 0.9771302236236454, + "grad_norm": 1.734375, + "learning_rate": 1.0230501594665121e-05, + "loss": 0.2104, + "step": 6740 + }, + { + "epoch": 0.978579971729912, + "grad_norm": 1.8828125, + "learning_rate": 1.0216004639025805e-05, + "loss": 0.2206, + "step": 6750 + }, + { + "epoch": 0.9800297198361785, + "grad_norm": 2.109375, + "learning_rate": 1.020150768338649e-05, + "loss": 0.1934, + "step": 6760 + }, + { + "epoch": 0.981479467942445, + "grad_norm": 2.046875, + "learning_rate": 1.0187010727747173e-05, + "loss": 0.2156, + "step": 6770 + }, + { + "epoch": 0.9829292160487115, + "grad_norm": 1.78125, + "learning_rate": 1.0172513772107859e-05, + "loss": 0.1732, + "step": 6780 + }, + { + "epoch": 0.984378964154978, + "grad_norm": 1.703125, + "learning_rate": 1.0158016816468543e-05, + "loss": 0.2188, + "step": 6790 + }, + { + "epoch": 0.9858287122612446, + "grad_norm": 1.7109375, + "learning_rate": 1.0143519860829227e-05, + "loss": 0.2172, + "step": 6800 + }, + { + "epoch": 0.9872784603675111, + "grad_norm": 1.765625, + "learning_rate": 1.0129022905189912e-05, + "loss": 0.2548, + "step": 6810 + }, + { + "epoch": 0.9887282084737777, + "grad_norm": 1.921875, + "learning_rate": 1.0114525949550595e-05, + "loss": 0.2045, + "step": 6820 + }, + { + "epoch": 0.9901779565800443, + "grad_norm": 1.7890625, + "learning_rate": 1.0100028993911279e-05, + "loss": 0.1958, + "step": 6830 + }, + { + "epoch": 0.9916277046863108, + "grad_norm": 1.859375, + "learning_rate": 1.0085532038271963e-05, + "loss": 0.2075, + "step": 6840 + }, + { + "epoch": 0.9930774527925773, + "grad_norm": 1.6015625, + "learning_rate": 1.0071035082632648e-05, + "loss": 0.1962, + "step": 6850 + }, + { + "epoch": 0.9945272008988438, + "grad_norm": 2.234375, + "learning_rate": 1.0056538126993332e-05, + "loss": 0.2256, + "step": 6860 + }, + { + "epoch": 0.9959769490051104, + "grad_norm": 1.8203125, + "learning_rate": 1.0042041171354015e-05, + "loss": 0.2345, + "step": 6870 + }, + { + "epoch": 0.9974266971113769, + "grad_norm": 1.65625, + "learning_rate": 1.0027544215714701e-05, + "loss": 0.2048, + "step": 6880 + }, + { + "epoch": 0.9988764452176434, + "grad_norm": 1.75, + "learning_rate": 1.0013047260075386e-05, + "loss": 0.1975, + "step": 6890 + }, + { + "epoch": 1.0002899496212534, + "grad_norm": 1.53125, + "learning_rate": 9.998550304436068e-06, + "loss": 0.1921, + "step": 6900 + }, + { + "epoch": 1.00173969772752, + "grad_norm": 1.4609375, + "learning_rate": 9.984053348796754e-06, + "loss": 0.1598, + "step": 6910 + }, + { + "epoch": 1.0031894458337864, + "grad_norm": 1.7734375, + "learning_rate": 9.969556393157439e-06, + "loss": 0.1326, + "step": 6920 + }, + { + "epoch": 1.004639193940053, + "grad_norm": 1.6015625, + "learning_rate": 9.955059437518121e-06, + "loss": 0.1603, + "step": 6930 + }, + { + "epoch": 1.0060889420463195, + "grad_norm": 1.5703125, + "learning_rate": 9.940562481878806e-06, + "loss": 0.1416, + "step": 6940 + }, + { + "epoch": 1.007538690152586, + "grad_norm": 1.59375, + "learning_rate": 9.92606552623949e-06, + "loss": 0.1626, + "step": 6950 + }, + { + "epoch": 1.0089884382588525, + "grad_norm": 2.03125, + "learning_rate": 9.911568570600175e-06, + "loss": 0.1731, + "step": 6960 + }, + { + "epoch": 1.010438186365119, + "grad_norm": 1.390625, + "learning_rate": 9.89707161496086e-06, + "loss": 0.1496, + "step": 6970 + }, + { + "epoch": 1.0118879344713856, + "grad_norm": 1.8046875, + "learning_rate": 9.882574659321544e-06, + "loss": 0.2041, + "step": 6980 + }, + { + "epoch": 1.013337682577652, + "grad_norm": 1.6171875, + "learning_rate": 9.868077703682228e-06, + "loss": 0.1567, + "step": 6990 + }, + { + "epoch": 1.0147874306839186, + "grad_norm": 1.171875, + "learning_rate": 9.85358074804291e-06, + "loss": 0.1432, + "step": 7000 + }, + { + "epoch": 1.0162371787901852, + "grad_norm": 1.5546875, + "learning_rate": 9.839083792403597e-06, + "loss": 0.1426, + "step": 7010 + }, + { + "epoch": 1.0176869268964517, + "grad_norm": 1.625, + "learning_rate": 9.824586836764281e-06, + "loss": 0.1792, + "step": 7020 + }, + { + "epoch": 1.0191366750027182, + "grad_norm": 1.421875, + "learning_rate": 9.810089881124964e-06, + "loss": 0.1704, + "step": 7030 + }, + { + "epoch": 1.0205864231089847, + "grad_norm": 1.28125, + "learning_rate": 9.795592925485648e-06, + "loss": 0.1448, + "step": 7040 + }, + { + "epoch": 1.0220361712152513, + "grad_norm": 1.7890625, + "learning_rate": 9.781095969846333e-06, + "loss": 0.1665, + "step": 7050 + }, + { + "epoch": 1.0234859193215178, + "grad_norm": 1.46875, + "learning_rate": 9.766599014207017e-06, + "loss": 0.1611, + "step": 7060 + }, + { + "epoch": 1.0249356674277845, + "grad_norm": 1.828125, + "learning_rate": 9.752102058567702e-06, + "loss": 0.147, + "step": 7070 + }, + { + "epoch": 1.026385415534051, + "grad_norm": 1.5546875, + "learning_rate": 9.737605102928386e-06, + "loss": 0.1515, + "step": 7080 + }, + { + "epoch": 1.0278351636403176, + "grad_norm": 1.6015625, + "learning_rate": 9.72310814728907e-06, + "loss": 0.1426, + "step": 7090 + }, + { + "epoch": 1.029284911746584, + "grad_norm": 1.9296875, + "learning_rate": 9.708611191649753e-06, + "loss": 0.1541, + "step": 7100 + }, + { + "epoch": 1.0307346598528506, + "grad_norm": 1.625, + "learning_rate": 9.69411423601044e-06, + "loss": 0.1467, + "step": 7110 + }, + { + "epoch": 1.0321844079591171, + "grad_norm": 1.8671875, + "learning_rate": 9.679617280371124e-06, + "loss": 0.1584, + "step": 7120 + }, + { + "epoch": 1.0336341560653837, + "grad_norm": 1.6484375, + "learning_rate": 9.665120324731806e-06, + "loss": 0.135, + "step": 7130 + }, + { + "epoch": 1.0350839041716502, + "grad_norm": 1.5234375, + "learning_rate": 9.650623369092491e-06, + "loss": 0.1903, + "step": 7140 + }, + { + "epoch": 1.0365336522779167, + "grad_norm": 1.4921875, + "learning_rate": 9.636126413453175e-06, + "loss": 0.1558, + "step": 7150 + }, + { + "epoch": 1.0379834003841832, + "grad_norm": 1.6015625, + "learning_rate": 9.62162945781386e-06, + "loss": 0.1407, + "step": 7160 + }, + { + "epoch": 1.0394331484904498, + "grad_norm": 1.9375, + "learning_rate": 9.607132502174544e-06, + "loss": 0.1712, + "step": 7170 + }, + { + "epoch": 1.0408828965967163, + "grad_norm": 1.46875, + "learning_rate": 9.592635546535229e-06, + "loss": 0.152, + "step": 7180 + }, + { + "epoch": 1.0423326447029828, + "grad_norm": 1.78125, + "learning_rate": 9.578138590895913e-06, + "loss": 0.1636, + "step": 7190 + }, + { + "epoch": 1.0437823928092493, + "grad_norm": 1.6953125, + "learning_rate": 9.563641635256596e-06, + "loss": 0.1673, + "step": 7200 + }, + { + "epoch": 1.0452321409155159, + "grad_norm": 1.8125, + "learning_rate": 9.549144679617282e-06, + "loss": 0.155, + "step": 7210 + }, + { + "epoch": 1.0466818890217824, + "grad_norm": 1.75, + "learning_rate": 9.534647723977966e-06, + "loss": 0.1487, + "step": 7220 + }, + { + "epoch": 1.048131637128049, + "grad_norm": 1.5078125, + "learning_rate": 9.520150768338649e-06, + "loss": 0.1465, + "step": 7230 + }, + { + "epoch": 1.0495813852343154, + "grad_norm": 1.734375, + "learning_rate": 9.505653812699333e-06, + "loss": 0.174, + "step": 7240 + }, + { + "epoch": 1.0510311333405822, + "grad_norm": 1.7109375, + "learning_rate": 9.491156857060018e-06, + "loss": 0.1594, + "step": 7250 + }, + { + "epoch": 1.0524808814468487, + "grad_norm": 1.5546875, + "learning_rate": 9.476659901420702e-06, + "loss": 0.1531, + "step": 7260 + }, + { + "epoch": 1.0539306295531152, + "grad_norm": 1.5859375, + "learning_rate": 9.462162945781387e-06, + "loss": 0.1459, + "step": 7270 + }, + { + "epoch": 1.0553803776593818, + "grad_norm": 1.09375, + "learning_rate": 9.447665990142071e-06, + "loss": 0.1616, + "step": 7280 + }, + { + "epoch": 1.0568301257656483, + "grad_norm": 1.296875, + "learning_rate": 9.433169034502755e-06, + "loss": 0.1244, + "step": 7290 + }, + { + "epoch": 1.0582798738719148, + "grad_norm": 1.5234375, + "learning_rate": 9.418672078863438e-06, + "loss": 0.1475, + "step": 7300 + }, + { + "epoch": 1.0597296219781813, + "grad_norm": 1.5703125, + "learning_rate": 9.404175123224124e-06, + "loss": 0.1433, + "step": 7310 + }, + { + "epoch": 1.0611793700844478, + "grad_norm": 1.421875, + "learning_rate": 9.389678167584809e-06, + "loss": 0.1458, + "step": 7320 + }, + { + "epoch": 1.0626291181907144, + "grad_norm": 1.7421875, + "learning_rate": 9.375181211945491e-06, + "loss": 0.1605, + "step": 7330 + }, + { + "epoch": 1.064078866296981, + "grad_norm": 1.8984375, + "learning_rate": 9.360684256306176e-06, + "loss": 0.1525, + "step": 7340 + }, + { + "epoch": 1.0655286144032474, + "grad_norm": 1.734375, + "learning_rate": 9.34618730066686e-06, + "loss": 0.1348, + "step": 7350 + }, + { + "epoch": 1.066978362509514, + "grad_norm": 1.4453125, + "learning_rate": 9.331690345027545e-06, + "loss": 0.145, + "step": 7360 + }, + { + "epoch": 1.0684281106157805, + "grad_norm": 1.734375, + "learning_rate": 9.317193389388229e-06, + "loss": 0.15, + "step": 7370 + }, + { + "epoch": 1.069877858722047, + "grad_norm": 1.5546875, + "learning_rate": 9.302696433748913e-06, + "loss": 0.1474, + "step": 7380 + }, + { + "epoch": 1.0713276068283135, + "grad_norm": 1.9296875, + "learning_rate": 9.288199478109598e-06, + "loss": 0.1575, + "step": 7390 + }, + { + "epoch": 1.07277735493458, + "grad_norm": 1.4375, + "learning_rate": 9.273702522470282e-06, + "loss": 0.1481, + "step": 7400 + }, + { + "epoch": 1.0742271030408466, + "grad_norm": 2.171875, + "learning_rate": 9.259205566830967e-06, + "loss": 0.1754, + "step": 7410 + }, + { + "epoch": 1.075676851147113, + "grad_norm": 1.53125, + "learning_rate": 9.244708611191651e-06, + "loss": 0.1724, + "step": 7420 + }, + { + "epoch": 1.0771265992533796, + "grad_norm": 1.3671875, + "learning_rate": 9.230211655552334e-06, + "loss": 0.1561, + "step": 7430 + }, + { + "epoch": 1.0785763473596464, + "grad_norm": 1.640625, + "learning_rate": 9.215714699913018e-06, + "loss": 0.1524, + "step": 7440 + }, + { + "epoch": 1.0800260954659129, + "grad_norm": 1.609375, + "learning_rate": 9.201217744273703e-06, + "loss": 0.1552, + "step": 7450 + }, + { + "epoch": 1.0814758435721794, + "grad_norm": 1.7578125, + "learning_rate": 9.186720788634387e-06, + "loss": 0.1723, + "step": 7460 + }, + { + "epoch": 1.082925591678446, + "grad_norm": 1.9453125, + "learning_rate": 9.172223832995072e-06, + "loss": 0.159, + "step": 7470 + }, + { + "epoch": 1.0843753397847125, + "grad_norm": 1.5390625, + "learning_rate": 9.157726877355756e-06, + "loss": 0.1582, + "step": 7480 + }, + { + "epoch": 1.085825087890979, + "grad_norm": 1.734375, + "learning_rate": 9.14322992171644e-06, + "loss": 0.1525, + "step": 7490 + }, + { + "epoch": 1.0872748359972455, + "grad_norm": 1.6640625, + "learning_rate": 9.128732966077125e-06, + "loss": 0.1524, + "step": 7500 + }, + { + "epoch": 1.088724584103512, + "grad_norm": 1.53125, + "learning_rate": 9.11423601043781e-06, + "loss": 0.1465, + "step": 7510 + }, + { + "epoch": 1.0901743322097786, + "grad_norm": 1.6953125, + "learning_rate": 9.099739054798494e-06, + "loss": 0.1521, + "step": 7520 + }, + { + "epoch": 1.091624080316045, + "grad_norm": 1.90625, + "learning_rate": 9.085242099159178e-06, + "loss": 0.1492, + "step": 7530 + }, + { + "epoch": 1.0930738284223116, + "grad_norm": 1.9140625, + "learning_rate": 9.07074514351986e-06, + "loss": 0.1351, + "step": 7540 + }, + { + "epoch": 1.0945235765285781, + "grad_norm": 1.5390625, + "learning_rate": 9.056248187880545e-06, + "loss": 0.1365, + "step": 7550 + }, + { + "epoch": 1.0959733246348446, + "grad_norm": 1.421875, + "learning_rate": 9.04175123224123e-06, + "loss": 0.1801, + "step": 7560 + }, + { + "epoch": 1.0974230727411112, + "grad_norm": 1.7421875, + "learning_rate": 9.027254276601914e-06, + "loss": 0.1233, + "step": 7570 + }, + { + "epoch": 1.0988728208473777, + "grad_norm": 1.4375, + "learning_rate": 9.012757320962598e-06, + "loss": 0.1306, + "step": 7580 + }, + { + "epoch": 1.1003225689536442, + "grad_norm": 1.6171875, + "learning_rate": 8.998260365323283e-06, + "loss": 0.1463, + "step": 7590 + }, + { + "epoch": 1.1017723170599107, + "grad_norm": 1.8828125, + "learning_rate": 8.983763409683967e-06, + "loss": 0.1423, + "step": 7600 + }, + { + "epoch": 1.1032220651661775, + "grad_norm": 1.5390625, + "learning_rate": 8.969266454044652e-06, + "loss": 0.1553, + "step": 7610 + }, + { + "epoch": 1.104671813272444, + "grad_norm": 1.5703125, + "learning_rate": 8.954769498405336e-06, + "loss": 0.1635, + "step": 7620 + }, + { + "epoch": 1.1061215613787105, + "grad_norm": 1.3671875, + "learning_rate": 8.94027254276602e-06, + "loss": 0.1437, + "step": 7630 + }, + { + "epoch": 1.107571309484977, + "grad_norm": 2.21875, + "learning_rate": 8.925775587126703e-06, + "loss": 0.1281, + "step": 7640 + }, + { + "epoch": 1.1090210575912436, + "grad_norm": 1.6328125, + "learning_rate": 8.91127863148739e-06, + "loss": 0.1542, + "step": 7650 + }, + { + "epoch": 1.11047080569751, + "grad_norm": 2.0625, + "learning_rate": 8.896781675848074e-06, + "loss": 0.1591, + "step": 7660 + }, + { + "epoch": 1.1119205538037766, + "grad_norm": 1.40625, + "learning_rate": 8.882284720208756e-06, + "loss": 0.1549, + "step": 7670 + }, + { + "epoch": 1.1133703019100432, + "grad_norm": 1.5546875, + "learning_rate": 8.867787764569441e-06, + "loss": 0.1521, + "step": 7680 + }, + { + "epoch": 1.1148200500163097, + "grad_norm": 2.25, + "learning_rate": 8.853290808930125e-06, + "loss": 0.1677, + "step": 7690 + }, + { + "epoch": 1.1162697981225762, + "grad_norm": 1.5859375, + "learning_rate": 8.83879385329081e-06, + "loss": 0.156, + "step": 7700 + }, + { + "epoch": 1.1177195462288427, + "grad_norm": 2.015625, + "learning_rate": 8.824296897651494e-06, + "loss": 0.1774, + "step": 7710 + }, + { + "epoch": 1.1191692943351093, + "grad_norm": 1.859375, + "learning_rate": 8.809799942012179e-06, + "loss": 0.146, + "step": 7720 + }, + { + "epoch": 1.1206190424413758, + "grad_norm": 2.109375, + "learning_rate": 8.795302986372863e-06, + "loss": 0.1492, + "step": 7730 + }, + { + "epoch": 1.1220687905476423, + "grad_norm": 1.7265625, + "learning_rate": 8.780806030733546e-06, + "loss": 0.1647, + "step": 7740 + }, + { + "epoch": 1.1235185386539088, + "grad_norm": 1.40625, + "learning_rate": 8.766309075094232e-06, + "loss": 0.1695, + "step": 7750 + }, + { + "epoch": 1.1249682867601754, + "grad_norm": 2.140625, + "learning_rate": 8.751812119454916e-06, + "loss": 0.152, + "step": 7760 + }, + { + "epoch": 1.1264180348664419, + "grad_norm": 1.8359375, + "learning_rate": 8.737315163815599e-06, + "loss": 0.1476, + "step": 7770 + }, + { + "epoch": 1.1278677829727084, + "grad_norm": 1.3125, + "learning_rate": 8.722818208176283e-06, + "loss": 0.1764, + "step": 7780 + }, + { + "epoch": 1.129317531078975, + "grad_norm": 1.578125, + "learning_rate": 8.708321252536968e-06, + "loss": 0.1698, + "step": 7790 + }, + { + "epoch": 1.1307672791852417, + "grad_norm": 2.15625, + "learning_rate": 8.693824296897652e-06, + "loss": 0.1433, + "step": 7800 + }, + { + "epoch": 1.1322170272915082, + "grad_norm": 2.1875, + "learning_rate": 8.679327341258337e-06, + "loss": 0.1604, + "step": 7810 + }, + { + "epoch": 1.1336667753977747, + "grad_norm": 1.4921875, + "learning_rate": 8.664830385619021e-06, + "loss": 0.1373, + "step": 7820 + }, + { + "epoch": 1.1351165235040412, + "grad_norm": 1.0234375, + "learning_rate": 8.650333429979705e-06, + "loss": 0.1478, + "step": 7830 + }, + { + "epoch": 1.1365662716103078, + "grad_norm": 1.4609375, + "learning_rate": 8.635836474340388e-06, + "loss": 0.1596, + "step": 7840 + }, + { + "epoch": 1.1380160197165743, + "grad_norm": 1.65625, + "learning_rate": 8.621339518701074e-06, + "loss": 0.142, + "step": 7850 + }, + { + "epoch": 1.1394657678228408, + "grad_norm": 1.8515625, + "learning_rate": 8.606842563061759e-06, + "loss": 0.1584, + "step": 7860 + }, + { + "epoch": 1.1409155159291073, + "grad_norm": 1.4609375, + "learning_rate": 8.592345607422441e-06, + "loss": 0.1359, + "step": 7870 + }, + { + "epoch": 1.1423652640353739, + "grad_norm": 1.15625, + "learning_rate": 8.577848651783126e-06, + "loss": 0.1619, + "step": 7880 + }, + { + "epoch": 1.1438150121416404, + "grad_norm": 1.5234375, + "learning_rate": 8.56335169614381e-06, + "loss": 0.167, + "step": 7890 + }, + { + "epoch": 1.145264760247907, + "grad_norm": 1.8359375, + "learning_rate": 8.548854740504495e-06, + "loss": 0.1569, + "step": 7900 + }, + { + "epoch": 1.1467145083541734, + "grad_norm": 1.78125, + "learning_rate": 8.534357784865179e-06, + "loss": 0.1508, + "step": 7910 + }, + { + "epoch": 1.14816425646044, + "grad_norm": 1.5859375, + "learning_rate": 8.519860829225864e-06, + "loss": 0.134, + "step": 7920 + }, + { + "epoch": 1.1496140045667065, + "grad_norm": 1.28125, + "learning_rate": 8.505363873586548e-06, + "loss": 0.1348, + "step": 7930 + }, + { + "epoch": 1.151063752672973, + "grad_norm": 1.578125, + "learning_rate": 8.49086691794723e-06, + "loss": 0.1487, + "step": 7940 + }, + { + "epoch": 1.1525135007792395, + "grad_norm": 1.734375, + "learning_rate": 8.476369962307917e-06, + "loss": 0.1224, + "step": 7950 + }, + { + "epoch": 1.153963248885506, + "grad_norm": 1.1328125, + "learning_rate": 8.461873006668601e-06, + "loss": 0.1374, + "step": 7960 + }, + { + "epoch": 1.1554129969917728, + "grad_norm": 1.15625, + "learning_rate": 8.447376051029284e-06, + "loss": 0.1422, + "step": 7970 + }, + { + "epoch": 1.156862745098039, + "grad_norm": 1.3203125, + "learning_rate": 8.432879095389968e-06, + "loss": 0.1474, + "step": 7980 + }, + { + "epoch": 1.1583124932043058, + "grad_norm": 1.5390625, + "learning_rate": 8.418382139750653e-06, + "loss": 0.1357, + "step": 7990 + }, + { + "epoch": 1.1597622413105724, + "grad_norm": 1.9140625, + "learning_rate": 8.403885184111337e-06, + "loss": 0.1493, + "step": 8000 + }, + { + "epoch": 1.161211989416839, + "grad_norm": 1.625, + "learning_rate": 8.389388228472022e-06, + "loss": 0.1421, + "step": 8010 + }, + { + "epoch": 1.1626617375231054, + "grad_norm": 1.5078125, + "learning_rate": 8.374891272832706e-06, + "loss": 0.1398, + "step": 8020 + }, + { + "epoch": 1.164111485629372, + "grad_norm": 1.546875, + "learning_rate": 8.36039431719339e-06, + "loss": 0.1532, + "step": 8030 + }, + { + "epoch": 1.1655612337356385, + "grad_norm": 1.84375, + "learning_rate": 8.345897361554073e-06, + "loss": 0.1476, + "step": 8040 + }, + { + "epoch": 1.167010981841905, + "grad_norm": 1.8359375, + "learning_rate": 8.33140040591476e-06, + "loss": 0.1678, + "step": 8050 + }, + { + "epoch": 1.1684607299481715, + "grad_norm": 1.4140625, + "learning_rate": 8.316903450275444e-06, + "loss": 0.136, + "step": 8060 + }, + { + "epoch": 1.169910478054438, + "grad_norm": 1.8359375, + "learning_rate": 8.302406494636126e-06, + "loss": 0.1572, + "step": 8070 + }, + { + "epoch": 1.1713602261607046, + "grad_norm": 1.65625, + "learning_rate": 8.28790953899681e-06, + "loss": 0.1598, + "step": 8080 + }, + { + "epoch": 1.172809974266971, + "grad_norm": 1.734375, + "learning_rate": 8.273412583357495e-06, + "loss": 0.1707, + "step": 8090 + }, + { + "epoch": 1.1742597223732376, + "grad_norm": 1.5390625, + "learning_rate": 8.25891562771818e-06, + "loss": 0.15, + "step": 8100 + }, + { + "epoch": 1.1757094704795041, + "grad_norm": 1.7421875, + "learning_rate": 8.244418672078864e-06, + "loss": 0.1266, + "step": 8110 + }, + { + "epoch": 1.1771592185857707, + "grad_norm": 2.03125, + "learning_rate": 8.229921716439548e-06, + "loss": 0.1601, + "step": 8120 + }, + { + "epoch": 1.1786089666920372, + "grad_norm": 1.7109375, + "learning_rate": 8.215424760800233e-06, + "loss": 0.1898, + "step": 8130 + }, + { + "epoch": 1.1800587147983037, + "grad_norm": 1.8203125, + "learning_rate": 8.200927805160917e-06, + "loss": 0.1515, + "step": 8140 + }, + { + "epoch": 1.1815084629045702, + "grad_norm": 2.109375, + "learning_rate": 8.186430849521602e-06, + "loss": 0.1474, + "step": 8150 + }, + { + "epoch": 1.182958211010837, + "grad_norm": 1.6796875, + "learning_rate": 8.171933893882286e-06, + "loss": 0.138, + "step": 8160 + }, + { + "epoch": 1.1844079591171033, + "grad_norm": 1.609375, + "learning_rate": 8.157436938242969e-06, + "loss": 0.1262, + "step": 8170 + }, + { + "epoch": 1.18585770722337, + "grad_norm": 1.9609375, + "learning_rate": 8.142939982603653e-06, + "loss": 0.116, + "step": 8180 + }, + { + "epoch": 1.1873074553296366, + "grad_norm": 2.140625, + "learning_rate": 8.128443026964338e-06, + "loss": 0.1985, + "step": 8190 + }, + { + "epoch": 1.188757203435903, + "grad_norm": 1.5546875, + "learning_rate": 8.113946071325022e-06, + "loss": 0.1349, + "step": 8200 + }, + { + "epoch": 1.1902069515421696, + "grad_norm": 1.5859375, + "learning_rate": 8.099449115685707e-06, + "loss": 0.1422, + "step": 8210 + }, + { + "epoch": 1.1916566996484361, + "grad_norm": 1.7578125, + "learning_rate": 8.084952160046391e-06, + "loss": 0.1342, + "step": 8220 + }, + { + "epoch": 1.1931064477547026, + "grad_norm": 1.4765625, + "learning_rate": 8.070455204407075e-06, + "loss": 0.1449, + "step": 8230 + }, + { + "epoch": 1.1945561958609692, + "grad_norm": 1.53125, + "learning_rate": 8.05595824876776e-06, + "loss": 0.155, + "step": 8240 + }, + { + "epoch": 1.1960059439672357, + "grad_norm": 1.546875, + "learning_rate": 8.041461293128444e-06, + "loss": 0.1173, + "step": 8250 + }, + { + "epoch": 1.1974556920735022, + "grad_norm": 1.4140625, + "learning_rate": 8.026964337489129e-06, + "loss": 0.1455, + "step": 8260 + }, + { + "epoch": 1.1989054401797687, + "grad_norm": 1.65625, + "learning_rate": 8.012467381849813e-06, + "loss": 0.1417, + "step": 8270 + }, + { + "epoch": 1.2003551882860353, + "grad_norm": 2.234375, + "learning_rate": 7.997970426210496e-06, + "loss": 0.128, + "step": 8280 + }, + { + "epoch": 1.2018049363923018, + "grad_norm": 2.21875, + "learning_rate": 7.98347347057118e-06, + "loss": 0.1469, + "step": 8290 + }, + { + "epoch": 1.2032546844985683, + "grad_norm": 1.4140625, + "learning_rate": 7.968976514931866e-06, + "loss": 0.1509, + "step": 8300 + }, + { + "epoch": 1.2047044326048348, + "grad_norm": 1.6796875, + "learning_rate": 7.954479559292549e-06, + "loss": 0.1482, + "step": 8310 + }, + { + "epoch": 1.2061541807111014, + "grad_norm": 1.953125, + "learning_rate": 7.939982603653233e-06, + "loss": 0.1689, + "step": 8320 + }, + { + "epoch": 1.2076039288173679, + "grad_norm": 2.109375, + "learning_rate": 7.925485648013918e-06, + "loss": 0.1507, + "step": 8330 + }, + { + "epoch": 1.2090536769236344, + "grad_norm": 1.671875, + "learning_rate": 7.910988692374602e-06, + "loss": 0.1567, + "step": 8340 + }, + { + "epoch": 1.2105034250299012, + "grad_norm": 1.3125, + "learning_rate": 7.896491736735287e-06, + "loss": 0.1494, + "step": 8350 + }, + { + "epoch": 1.2119531731361677, + "grad_norm": 1.5390625, + "learning_rate": 7.881994781095971e-06, + "loss": 0.1451, + "step": 8360 + }, + { + "epoch": 1.2134029212424342, + "grad_norm": 1.6328125, + "learning_rate": 7.867497825456655e-06, + "loss": 0.1377, + "step": 8370 + }, + { + "epoch": 1.2148526693487007, + "grad_norm": 0.95703125, + "learning_rate": 7.853000869817338e-06, + "loss": 0.1404, + "step": 8380 + }, + { + "epoch": 1.2163024174549673, + "grad_norm": 1.859375, + "learning_rate": 7.838503914178023e-06, + "loss": 0.1345, + "step": 8390 + }, + { + "epoch": 1.2177521655612338, + "grad_norm": 1.4921875, + "learning_rate": 7.824006958538709e-06, + "loss": 0.1476, + "step": 8400 + }, + { + "epoch": 1.2192019136675003, + "grad_norm": 1.5078125, + "learning_rate": 7.809510002899391e-06, + "loss": 0.1515, + "step": 8410 + }, + { + "epoch": 1.2206516617737668, + "grad_norm": 1.484375, + "learning_rate": 7.795013047260076e-06, + "loss": 0.1585, + "step": 8420 + }, + { + "epoch": 1.2221014098800334, + "grad_norm": 1.3359375, + "learning_rate": 7.78051609162076e-06, + "loss": 0.1607, + "step": 8430 + }, + { + "epoch": 1.2235511579862999, + "grad_norm": 1.625, + "learning_rate": 7.766019135981445e-06, + "loss": 0.1474, + "step": 8440 + }, + { + "epoch": 1.2250009060925664, + "grad_norm": 1.625, + "learning_rate": 7.751522180342129e-06, + "loss": 0.1599, + "step": 8450 + }, + { + "epoch": 1.226450654198833, + "grad_norm": 2.40625, + "learning_rate": 7.737025224702814e-06, + "loss": 0.1479, + "step": 8460 + }, + { + "epoch": 1.2279004023050994, + "grad_norm": 1.328125, + "learning_rate": 7.722528269063498e-06, + "loss": 0.1627, + "step": 8470 + }, + { + "epoch": 1.229350150411366, + "grad_norm": 1.5859375, + "learning_rate": 7.70803131342418e-06, + "loss": 0.1603, + "step": 8480 + }, + { + "epoch": 1.2307998985176325, + "grad_norm": 1.8125, + "learning_rate": 7.693534357784865e-06, + "loss": 0.1657, + "step": 8490 + }, + { + "epoch": 1.232249646623899, + "grad_norm": 1.328125, + "learning_rate": 7.679037402145551e-06, + "loss": 0.119, + "step": 8500 + }, + { + "epoch": 1.2336993947301655, + "grad_norm": 2.203125, + "learning_rate": 7.664540446506234e-06, + "loss": 0.1333, + "step": 8510 + }, + { + "epoch": 1.2351491428364323, + "grad_norm": 1.59375, + "learning_rate": 7.650043490866918e-06, + "loss": 0.138, + "step": 8520 + }, + { + "epoch": 1.2365988909426986, + "grad_norm": 1.328125, + "learning_rate": 7.635546535227603e-06, + "loss": 0.1264, + "step": 8530 + }, + { + "epoch": 1.2380486390489653, + "grad_norm": 1.1484375, + "learning_rate": 7.621049579588286e-06, + "loss": 0.1392, + "step": 8540 + }, + { + "epoch": 1.2394983871552319, + "grad_norm": 1.7109375, + "learning_rate": 7.606552623948972e-06, + "loss": 0.1424, + "step": 8550 + }, + { + "epoch": 1.2409481352614984, + "grad_norm": 1.2890625, + "learning_rate": 7.592055668309656e-06, + "loss": 0.1755, + "step": 8560 + }, + { + "epoch": 1.242397883367765, + "grad_norm": 1.9296875, + "learning_rate": 7.57755871267034e-06, + "loss": 0.1673, + "step": 8570 + }, + { + "epoch": 1.2438476314740314, + "grad_norm": 1.734375, + "learning_rate": 7.563061757031024e-06, + "loss": 0.1711, + "step": 8580 + }, + { + "epoch": 1.245297379580298, + "grad_norm": 1.65625, + "learning_rate": 7.5485648013917076e-06, + "loss": 0.1388, + "step": 8590 + }, + { + "epoch": 1.2467471276865645, + "grad_norm": 1.484375, + "learning_rate": 7.534067845752393e-06, + "loss": 0.1503, + "step": 8600 + }, + { + "epoch": 1.248196875792831, + "grad_norm": 1.796875, + "learning_rate": 7.519570890113077e-06, + "loss": 0.1332, + "step": 8610 + }, + { + "epoch": 1.2496466238990975, + "grad_norm": 1.625, + "learning_rate": 7.505073934473761e-06, + "loss": 0.1447, + "step": 8620 + }, + { + "epoch": 1.251096372005364, + "grad_norm": 1.2578125, + "learning_rate": 7.490576978834445e-06, + "loss": 0.1505, + "step": 8630 + }, + { + "epoch": 1.2525461201116306, + "grad_norm": 2.0, + "learning_rate": 7.476080023195129e-06, + "loss": 0.156, + "step": 8640 + }, + { + "epoch": 1.253995868217897, + "grad_norm": 1.5234375, + "learning_rate": 7.461583067555814e-06, + "loss": 0.1516, + "step": 8650 + }, + { + "epoch": 1.2554456163241636, + "grad_norm": 1.8359375, + "learning_rate": 7.4470861119164985e-06, + "loss": 0.1396, + "step": 8660 + }, + { + "epoch": 1.2568953644304302, + "grad_norm": 1.6640625, + "learning_rate": 7.432589156277182e-06, + "loss": 0.122, + "step": 8670 + }, + { + "epoch": 1.2583451125366967, + "grad_norm": 1.3046875, + "learning_rate": 7.4180922006378665e-06, + "loss": 0.1394, + "step": 8680 + }, + { + "epoch": 1.2597948606429634, + "grad_norm": 1.4609375, + "learning_rate": 7.40359524499855e-06, + "loss": 0.133, + "step": 8690 + }, + { + "epoch": 1.2612446087492297, + "grad_norm": 1.6328125, + "learning_rate": 7.389098289359235e-06, + "loss": 0.148, + "step": 8700 + }, + { + "epoch": 1.2626943568554965, + "grad_norm": 1.8046875, + "learning_rate": 7.37460133371992e-06, + "loss": 0.1508, + "step": 8710 + }, + { + "epoch": 1.2641441049617628, + "grad_norm": 2.25, + "learning_rate": 7.360104378080603e-06, + "loss": 0.1478, + "step": 8720 + }, + { + "epoch": 1.2655938530680295, + "grad_norm": 1.6171875, + "learning_rate": 7.345607422441288e-06, + "loss": 0.1365, + "step": 8730 + }, + { + "epoch": 1.267043601174296, + "grad_norm": 1.203125, + "learning_rate": 7.331110466801971e-06, + "loss": 0.1482, + "step": 8740 + }, + { + "epoch": 1.2684933492805626, + "grad_norm": 1.5625, + "learning_rate": 7.3166135111626565e-06, + "loss": 0.1286, + "step": 8750 + }, + { + "epoch": 1.269943097386829, + "grad_norm": 1.3515625, + "learning_rate": 7.302116555523341e-06, + "loss": 0.146, + "step": 8760 + }, + { + "epoch": 1.2713928454930956, + "grad_norm": 1.65625, + "learning_rate": 7.2876195998840245e-06, + "loss": 0.1333, + "step": 8770 + }, + { + "epoch": 1.2728425935993621, + "grad_norm": 1.5390625, + "learning_rate": 7.273122644244709e-06, + "loss": 0.1275, + "step": 8780 + }, + { + "epoch": 1.2742923417056287, + "grad_norm": 1.796875, + "learning_rate": 7.258625688605393e-06, + "loss": 0.1324, + "step": 8790 + }, + { + "epoch": 1.2757420898118952, + "grad_norm": 1.2578125, + "learning_rate": 7.244128732966078e-06, + "loss": 0.1355, + "step": 8800 + }, + { + "epoch": 1.2771918379181617, + "grad_norm": 1.328125, + "learning_rate": 7.229631777326762e-06, + "loss": 0.1194, + "step": 8810 + }, + { + "epoch": 1.2786415860244282, + "grad_norm": 1.859375, + "learning_rate": 7.215134821687446e-06, + "loss": 0.1591, + "step": 8820 + }, + { + "epoch": 1.2800913341306948, + "grad_norm": 2.125, + "learning_rate": 7.20063786604813e-06, + "loss": 0.1346, + "step": 8830 + }, + { + "epoch": 1.2815410822369613, + "grad_norm": 1.59375, + "learning_rate": 7.186140910408815e-06, + "loss": 0.1511, + "step": 8840 + }, + { + "epoch": 1.2829908303432278, + "grad_norm": 2.4375, + "learning_rate": 7.171643954769499e-06, + "loss": 0.1494, + "step": 8850 + }, + { + "epoch": 1.2844405784494943, + "grad_norm": 1.3203125, + "learning_rate": 7.1571469991301834e-06, + "loss": 0.1316, + "step": 8860 + }, + { + "epoch": 1.2858903265557609, + "grad_norm": 1.7421875, + "learning_rate": 7.142650043490868e-06, + "loss": 0.1379, + "step": 8870 + }, + { + "epoch": 1.2873400746620276, + "grad_norm": 1.2421875, + "learning_rate": 7.128153087851551e-06, + "loss": 0.1206, + "step": 8880 + }, + { + "epoch": 1.288789822768294, + "grad_norm": 1.4375, + "learning_rate": 7.113656132212236e-06, + "loss": 0.1305, + "step": 8890 + }, + { + "epoch": 1.2902395708745606, + "grad_norm": 1.859375, + "learning_rate": 7.09915917657292e-06, + "loss": 0.1371, + "step": 8900 + }, + { + "epoch": 1.291689318980827, + "grad_norm": 1.2578125, + "learning_rate": 7.084662220933605e-06, + "loss": 0.1547, + "step": 8910 + }, + { + "epoch": 1.2931390670870937, + "grad_norm": 1.6640625, + "learning_rate": 7.070165265294289e-06, + "loss": 0.1494, + "step": 8920 + }, + { + "epoch": 1.2945888151933602, + "grad_norm": 1.53125, + "learning_rate": 7.055668309654973e-06, + "loss": 0.1254, + "step": 8930 + }, + { + "epoch": 1.2960385632996267, + "grad_norm": 1.8125, + "learning_rate": 7.041171354015657e-06, + "loss": 0.1467, + "step": 8940 + }, + { + "epoch": 1.2974883114058933, + "grad_norm": 2.1875, + "learning_rate": 7.026674398376342e-06, + "loss": 0.181, + "step": 8950 + }, + { + "epoch": 1.2989380595121598, + "grad_norm": 1.234375, + "learning_rate": 7.012177442737026e-06, + "loss": 0.1649, + "step": 8960 + }, + { + "epoch": 1.3003878076184263, + "grad_norm": 1.5703125, + "learning_rate": 6.99768048709771e-06, + "loss": 0.1485, + "step": 8970 + }, + { + "epoch": 1.3018375557246928, + "grad_norm": 1.578125, + "learning_rate": 6.983183531458394e-06, + "loss": 0.1288, + "step": 8980 + }, + { + "epoch": 1.3032873038309594, + "grad_norm": 1.84375, + "learning_rate": 6.968686575819078e-06, + "loss": 0.1448, + "step": 8990 + }, + { + "epoch": 1.304737051937226, + "grad_norm": 1.8359375, + "learning_rate": 6.9541896201797636e-06, + "loss": 0.1504, + "step": 9000 + }, + { + "epoch": 1.3061868000434924, + "grad_norm": 1.765625, + "learning_rate": 6.939692664540447e-06, + "loss": 0.1007, + "step": 9010 + }, + { + "epoch": 1.307636548149759, + "grad_norm": 1.4140625, + "learning_rate": 6.9251957089011316e-06, + "loss": 0.1324, + "step": 9020 + }, + { + "epoch": 1.3090862962560255, + "grad_norm": 1.078125, + "learning_rate": 6.910698753261815e-06, + "loss": 0.1452, + "step": 9030 + }, + { + "epoch": 1.310536044362292, + "grad_norm": 1.34375, + "learning_rate": 6.8962017976224995e-06, + "loss": 0.1224, + "step": 9040 + }, + { + "epoch": 1.3119857924685585, + "grad_norm": 1.2890625, + "learning_rate": 6.881704841983185e-06, + "loss": 0.1559, + "step": 9050 + }, + { + "epoch": 1.313435540574825, + "grad_norm": 1.7109375, + "learning_rate": 6.867207886343868e-06, + "loss": 0.151, + "step": 9060 + }, + { + "epoch": 1.3148852886810918, + "grad_norm": 1.765625, + "learning_rate": 6.852710930704553e-06, + "loss": 0.1558, + "step": 9070 + }, + { + "epoch": 1.316335036787358, + "grad_norm": 1.484375, + "learning_rate": 6.838213975065236e-06, + "loss": 0.137, + "step": 9080 + }, + { + "epoch": 1.3177847848936248, + "grad_norm": 1.90625, + "learning_rate": 6.823717019425921e-06, + "loss": 0.1493, + "step": 9090 + }, + { + "epoch": 1.3192345329998914, + "grad_norm": 1.3359375, + "learning_rate": 6.809220063786606e-06, + "loss": 0.136, + "step": 9100 + }, + { + "epoch": 1.3206842811061579, + "grad_norm": 1.6640625, + "learning_rate": 6.79472310814729e-06, + "loss": 0.156, + "step": 9110 + }, + { + "epoch": 1.3221340292124244, + "grad_norm": 1.265625, + "learning_rate": 6.780226152507974e-06, + "loss": 0.1402, + "step": 9120 + }, + { + "epoch": 1.323583777318691, + "grad_norm": 1.9375, + "learning_rate": 6.765729196868658e-06, + "loss": 0.1521, + "step": 9130 + }, + { + "epoch": 1.3250335254249574, + "grad_norm": 1.859375, + "learning_rate": 6.751232241229342e-06, + "loss": 0.1757, + "step": 9140 + }, + { + "epoch": 1.326483273531224, + "grad_norm": 1.5078125, + "learning_rate": 6.736735285590027e-06, + "loss": 0.13, + "step": 9150 + }, + { + "epoch": 1.3279330216374905, + "grad_norm": 1.4453125, + "learning_rate": 6.722238329950711e-06, + "loss": 0.1636, + "step": 9160 + }, + { + "epoch": 1.329382769743757, + "grad_norm": 1.65625, + "learning_rate": 6.707741374311395e-06, + "loss": 0.1518, + "step": 9170 + }, + { + "epoch": 1.3308325178500235, + "grad_norm": 1.5625, + "learning_rate": 6.693244418672079e-06, + "loss": 0.131, + "step": 9180 + }, + { + "epoch": 1.33228226595629, + "grad_norm": 1.6953125, + "learning_rate": 6.678747463032763e-06, + "loss": 0.1467, + "step": 9190 + }, + { + "epoch": 1.3337320140625566, + "grad_norm": 1.4765625, + "learning_rate": 6.6642505073934485e-06, + "loss": 0.142, + "step": 9200 + }, + { + "epoch": 1.3351817621688231, + "grad_norm": 2.125, + "learning_rate": 6.649753551754132e-06, + "loss": 0.1569, + "step": 9210 + }, + { + "epoch": 1.3366315102750896, + "grad_norm": 1.59375, + "learning_rate": 6.6352565961148165e-06, + "loss": 0.1396, + "step": 9220 + }, + { + "epoch": 1.3380812583813562, + "grad_norm": 1.828125, + "learning_rate": 6.6207596404755e-06, + "loss": 0.1422, + "step": 9230 + }, + { + "epoch": 1.339531006487623, + "grad_norm": 1.421875, + "learning_rate": 6.606262684836185e-06, + "loss": 0.1591, + "step": 9240 + }, + { + "epoch": 1.3409807545938892, + "grad_norm": 1.6796875, + "learning_rate": 6.59176572919687e-06, + "loss": 0.1591, + "step": 9250 + }, + { + "epoch": 1.342430502700156, + "grad_norm": 1.6015625, + "learning_rate": 6.577268773557553e-06, + "loss": 0.1325, + "step": 9260 + }, + { + "epoch": 1.3438802508064223, + "grad_norm": 1.6171875, + "learning_rate": 6.562771817918238e-06, + "loss": 0.1642, + "step": 9270 + }, + { + "epoch": 1.345329998912689, + "grad_norm": 1.765625, + "learning_rate": 6.548274862278921e-06, + "loss": 0.1454, + "step": 9280 + }, + { + "epoch": 1.3467797470189555, + "grad_norm": 1.8671875, + "learning_rate": 6.5337779066396066e-06, + "loss": 0.1538, + "step": 9290 + }, + { + "epoch": 1.348229495125222, + "grad_norm": 2.25, + "learning_rate": 6.519280951000291e-06, + "loss": 0.1572, + "step": 9300 + }, + { + "epoch": 1.3496792432314886, + "grad_norm": 1.703125, + "learning_rate": 6.5047839953609746e-06, + "loss": 0.1592, + "step": 9310 + }, + { + "epoch": 1.351128991337755, + "grad_norm": 1.421875, + "learning_rate": 6.490287039721659e-06, + "loss": 0.1377, + "step": 9320 + }, + { + "epoch": 1.3525787394440216, + "grad_norm": 1.421875, + "learning_rate": 6.4757900840823425e-06, + "loss": 0.1332, + "step": 9330 + }, + { + "epoch": 1.3540284875502882, + "grad_norm": 1.625, + "learning_rate": 6.461293128443028e-06, + "loss": 0.1433, + "step": 9340 + }, + { + "epoch": 1.3554782356565547, + "grad_norm": 1.1484375, + "learning_rate": 6.446796172803712e-06, + "loss": 0.1454, + "step": 9350 + }, + { + "epoch": 1.3569279837628212, + "grad_norm": 2.1875, + "learning_rate": 6.432299217164396e-06, + "loss": 0.1461, + "step": 9360 + }, + { + "epoch": 1.3583777318690877, + "grad_norm": 2.171875, + "learning_rate": 6.41780226152508e-06, + "loss": 0.1394, + "step": 9370 + }, + { + "epoch": 1.3598274799753542, + "grad_norm": 1.0234375, + "learning_rate": 6.403305305885764e-06, + "loss": 0.1265, + "step": 9380 + }, + { + "epoch": 1.3612772280816208, + "grad_norm": 1.625, + "learning_rate": 6.388808350246449e-06, + "loss": 0.1522, + "step": 9390 + }, + { + "epoch": 1.3627269761878873, + "grad_norm": 1.640625, + "learning_rate": 6.3743113946071335e-06, + "loss": 0.1219, + "step": 9400 + }, + { + "epoch": 1.3641767242941538, + "grad_norm": 1.375, + "learning_rate": 6.359814438967817e-06, + "loss": 0.128, + "step": 9410 + }, + { + "epoch": 1.3656264724004203, + "grad_norm": 1.6328125, + "learning_rate": 6.3453174833285014e-06, + "loss": 0.137, + "step": 9420 + }, + { + "epoch": 1.367076220506687, + "grad_norm": 1.359375, + "learning_rate": 6.330820527689185e-06, + "loss": 0.1358, + "step": 9430 + }, + { + "epoch": 1.3685259686129534, + "grad_norm": 2.28125, + "learning_rate": 6.31632357204987e-06, + "loss": 0.1379, + "step": 9440 + }, + { + "epoch": 1.3699757167192201, + "grad_norm": 2.265625, + "learning_rate": 6.301826616410555e-06, + "loss": 0.1302, + "step": 9450 + }, + { + "epoch": 1.3714254648254864, + "grad_norm": 1.8828125, + "learning_rate": 6.287329660771238e-06, + "loss": 0.1253, + "step": 9460 + }, + { + "epoch": 1.3728752129317532, + "grad_norm": 1.5234375, + "learning_rate": 6.272832705131923e-06, + "loss": 0.1174, + "step": 9470 + }, + { + "epoch": 1.3743249610380197, + "grad_norm": 1.765625, + "learning_rate": 6.258335749492607e-06, + "loss": 0.1542, + "step": 9480 + }, + { + "epoch": 1.3757747091442862, + "grad_norm": 2.03125, + "learning_rate": 6.2438387938532915e-06, + "loss": 0.1462, + "step": 9490 + }, + { + "epoch": 1.3772244572505528, + "grad_norm": 2.09375, + "learning_rate": 6.229341838213976e-06, + "loss": 0.135, + "step": 9500 + }, + { + "epoch": 1.3786742053568193, + "grad_norm": 1.5078125, + "learning_rate": 6.2148448825746595e-06, + "loss": 0.1356, + "step": 9510 + }, + { + "epoch": 1.3801239534630858, + "grad_norm": 1.875, + "learning_rate": 6.200347926935344e-06, + "loss": 0.1441, + "step": 9520 + }, + { + "epoch": 1.3815737015693523, + "grad_norm": 1.21875, + "learning_rate": 6.185850971296028e-06, + "loss": 0.1229, + "step": 9530 + }, + { + "epoch": 1.3830234496756189, + "grad_norm": 1.9921875, + "learning_rate": 6.171354015656713e-06, + "loss": 0.144, + "step": 9540 + }, + { + "epoch": 1.3844731977818854, + "grad_norm": 2.03125, + "learning_rate": 6.156857060017397e-06, + "loss": 0.17, + "step": 9550 + }, + { + "epoch": 1.385922945888152, + "grad_norm": 1.4765625, + "learning_rate": 6.142360104378082e-06, + "loss": 0.1334, + "step": 9560 + }, + { + "epoch": 1.3873726939944184, + "grad_norm": 1.8125, + "learning_rate": 6.127863148738765e-06, + "loss": 0.1346, + "step": 9570 + }, + { + "epoch": 1.388822442100685, + "grad_norm": 1.828125, + "learning_rate": 6.1133661930994496e-06, + "loss": 0.1415, + "step": 9580 + }, + { + "epoch": 1.3902721902069515, + "grad_norm": 1.9140625, + "learning_rate": 6.098869237460134e-06, + "loss": 0.1272, + "step": 9590 + }, + { + "epoch": 1.391721938313218, + "grad_norm": 1.625, + "learning_rate": 6.084372281820818e-06, + "loss": 0.15, + "step": 9600 + }, + { + "epoch": 1.3931716864194845, + "grad_norm": 1.59375, + "learning_rate": 6.069875326181503e-06, + "loss": 0.1199, + "step": 9610 + }, + { + "epoch": 1.3946214345257513, + "grad_norm": 1.4609375, + "learning_rate": 6.055378370542186e-06, + "loss": 0.1472, + "step": 9620 + }, + { + "epoch": 1.3960711826320176, + "grad_norm": 1.1328125, + "learning_rate": 6.040881414902871e-06, + "loss": 0.142, + "step": 9630 + }, + { + "epoch": 1.3975209307382843, + "grad_norm": 1.2421875, + "learning_rate": 6.026384459263555e-06, + "loss": 0.1429, + "step": 9640 + }, + { + "epoch": 1.3989706788445508, + "grad_norm": 1.984375, + "learning_rate": 6.01188750362424e-06, + "loss": 0.1489, + "step": 9650 + }, + { + "epoch": 1.4004204269508174, + "grad_norm": 1.6953125, + "learning_rate": 5.997390547984924e-06, + "loss": 0.1257, + "step": 9660 + }, + { + "epoch": 1.401870175057084, + "grad_norm": 1.7421875, + "learning_rate": 5.982893592345608e-06, + "loss": 0.1153, + "step": 9670 + }, + { + "epoch": 1.4033199231633504, + "grad_norm": 1.8515625, + "learning_rate": 5.968396636706292e-06, + "loss": 0.1309, + "step": 9680 + }, + { + "epoch": 1.404769671269617, + "grad_norm": 0.96875, + "learning_rate": 5.953899681066977e-06, + "loss": 0.1321, + "step": 9690 + }, + { + "epoch": 1.4062194193758835, + "grad_norm": 2.0625, + "learning_rate": 5.939402725427661e-06, + "loss": 0.1538, + "step": 9700 + }, + { + "epoch": 1.40766916748215, + "grad_norm": 1.4921875, + "learning_rate": 5.924905769788345e-06, + "loss": 0.1275, + "step": 9710 + }, + { + "epoch": 1.4091189155884165, + "grad_norm": 1.265625, + "learning_rate": 5.910408814149029e-06, + "loss": 0.1357, + "step": 9720 + }, + { + "epoch": 1.410568663694683, + "grad_norm": 1.703125, + "learning_rate": 5.895911858509713e-06, + "loss": 0.1476, + "step": 9730 + }, + { + "epoch": 1.4120184118009496, + "grad_norm": 1.75, + "learning_rate": 5.8814149028703985e-06, + "loss": 0.1362, + "step": 9740 + }, + { + "epoch": 1.413468159907216, + "grad_norm": 1.6484375, + "learning_rate": 5.866917947231082e-06, + "loss": 0.1374, + "step": 9750 + }, + { + "epoch": 1.4149179080134826, + "grad_norm": 2.1875, + "learning_rate": 5.8524209915917665e-06, + "loss": 0.1203, + "step": 9760 + }, + { + "epoch": 1.4163676561197491, + "grad_norm": 1.5234375, + "learning_rate": 5.83792403595245e-06, + "loss": 0.1219, + "step": 9770 + }, + { + "epoch": 1.4178174042260157, + "grad_norm": 1.8046875, + "learning_rate": 5.8234270803131345e-06, + "loss": 0.1472, + "step": 9780 + }, + { + "epoch": 1.4192671523322824, + "grad_norm": 1.1484375, + "learning_rate": 5.80893012467382e-06, + "loss": 0.1456, + "step": 9790 + }, + { + "epoch": 1.4207169004385487, + "grad_norm": 2.28125, + "learning_rate": 5.794433169034503e-06, + "loss": 0.1574, + "step": 9800 + }, + { + "epoch": 1.4221666485448154, + "grad_norm": 1.9296875, + "learning_rate": 5.779936213395188e-06, + "loss": 0.1487, + "step": 9810 + }, + { + "epoch": 1.4236163966510818, + "grad_norm": 1.3984375, + "learning_rate": 5.765439257755871e-06, + "loss": 0.1461, + "step": 9820 + }, + { + "epoch": 1.4250661447573485, + "grad_norm": 2.078125, + "learning_rate": 5.750942302116556e-06, + "loss": 0.1235, + "step": 9830 + }, + { + "epoch": 1.426515892863615, + "grad_norm": 2.1875, + "learning_rate": 5.736445346477241e-06, + "loss": 0.1512, + "step": 9840 + }, + { + "epoch": 1.4279656409698815, + "grad_norm": 1.5546875, + "learning_rate": 5.721948390837925e-06, + "loss": 0.1426, + "step": 9850 + }, + { + "epoch": 1.429415389076148, + "grad_norm": 2.5625, + "learning_rate": 5.707451435198609e-06, + "loss": 0.1434, + "step": 9860 + }, + { + "epoch": 1.4308651371824146, + "grad_norm": 2.046875, + "learning_rate": 5.6929544795592926e-06, + "loss": 0.1647, + "step": 9870 + }, + { + "epoch": 1.4323148852886811, + "grad_norm": 1.0078125, + "learning_rate": 5.678457523919977e-06, + "loss": 0.138, + "step": 9880 + }, + { + "epoch": 1.4337646333949476, + "grad_norm": 1.34375, + "learning_rate": 5.663960568280662e-06, + "loss": 0.1188, + "step": 9890 + }, + { + "epoch": 1.4352143815012142, + "grad_norm": 1.7109375, + "learning_rate": 5.649463612641346e-06, + "loss": 0.1315, + "step": 9900 + }, + { + "epoch": 1.4366641296074807, + "grad_norm": 1.8203125, + "learning_rate": 5.63496665700203e-06, + "loss": 0.1606, + "step": 9910 + }, + { + "epoch": 1.4381138777137472, + "grad_norm": 1.5234375, + "learning_rate": 5.620469701362714e-06, + "loss": 0.1469, + "step": 9920 + }, + { + "epoch": 1.4395636258200137, + "grad_norm": 1.4375, + "learning_rate": 5.605972745723398e-06, + "loss": 0.1186, + "step": 9930 + }, + { + "epoch": 1.4410133739262803, + "grad_norm": 1.9765625, + "learning_rate": 5.5914757900840835e-06, + "loss": 0.1299, + "step": 9940 + }, + { + "epoch": 1.4424631220325468, + "grad_norm": 1.484375, + "learning_rate": 5.576978834444767e-06, + "loss": 0.1372, + "step": 9950 + }, + { + "epoch": 1.4439128701388133, + "grad_norm": 1.296875, + "learning_rate": 5.5624818788054515e-06, + "loss": 0.1329, + "step": 9960 + }, + { + "epoch": 1.4453626182450798, + "grad_norm": 1.9375, + "learning_rate": 5.547984923166135e-06, + "loss": 0.1148, + "step": 9970 + }, + { + "epoch": 1.4468123663513466, + "grad_norm": 1.8671875, + "learning_rate": 5.5334879675268195e-06, + "loss": 0.1321, + "step": 9980 + }, + { + "epoch": 1.4482621144576129, + "grad_norm": 1.9296875, + "learning_rate": 5.518991011887505e-06, + "loss": 0.1128, + "step": 9990 + }, + { + "epoch": 1.4497118625638796, + "grad_norm": 1.8671875, + "learning_rate": 5.504494056248188e-06, + "loss": 0.1225, + "step": 10000 + }, + { + "epoch": 1.451161610670146, + "grad_norm": 1.5703125, + "learning_rate": 5.489997100608873e-06, + "loss": 0.1437, + "step": 10010 + }, + { + "epoch": 1.4526113587764127, + "grad_norm": 1.8984375, + "learning_rate": 5.475500144969556e-06, + "loss": 0.1649, + "step": 10020 + }, + { + "epoch": 1.4540611068826792, + "grad_norm": 1.71875, + "learning_rate": 5.461003189330241e-06, + "loss": 0.1737, + "step": 10030 + }, + { + "epoch": 1.4555108549889457, + "grad_norm": 1.90625, + "learning_rate": 5.446506233690926e-06, + "loss": 0.1581, + "step": 10040 + }, + { + "epoch": 1.4569606030952122, + "grad_norm": 1.515625, + "learning_rate": 5.4320092780516095e-06, + "loss": 0.1316, + "step": 10050 + }, + { + "epoch": 1.4584103512014788, + "grad_norm": 0.97265625, + "learning_rate": 5.417512322412294e-06, + "loss": 0.138, + "step": 10060 + }, + { + "epoch": 1.4598600993077453, + "grad_norm": 2.046875, + "learning_rate": 5.4030153667729775e-06, + "loss": 0.1498, + "step": 10070 + }, + { + "epoch": 1.4613098474140118, + "grad_norm": 1.3359375, + "learning_rate": 5.388518411133662e-06, + "loss": 0.1246, + "step": 10080 + }, + { + "epoch": 1.4627595955202783, + "grad_norm": 1.640625, + "learning_rate": 5.374021455494347e-06, + "loss": 0.1495, + "step": 10090 + }, + { + "epoch": 1.4642093436265449, + "grad_norm": 1.578125, + "learning_rate": 5.359524499855031e-06, + "loss": 0.127, + "step": 10100 + }, + { + "epoch": 1.4656590917328114, + "grad_norm": 2.28125, + "learning_rate": 5.345027544215715e-06, + "loss": 0.1651, + "step": 10110 + }, + { + "epoch": 1.467108839839078, + "grad_norm": 1.8203125, + "learning_rate": 5.330530588576399e-06, + "loss": 0.1323, + "step": 10120 + }, + { + "epoch": 1.4685585879453444, + "grad_norm": 2.1875, + "learning_rate": 5.316033632937083e-06, + "loss": 0.1609, + "step": 10130 + }, + { + "epoch": 1.470008336051611, + "grad_norm": 1.78125, + "learning_rate": 5.3015366772977684e-06, + "loss": 0.1559, + "step": 10140 + }, + { + "epoch": 1.4714580841578775, + "grad_norm": 2.1875, + "learning_rate": 5.287039721658452e-06, + "loss": 0.1457, + "step": 10150 + }, + { + "epoch": 1.472907832264144, + "grad_norm": 0.97265625, + "learning_rate": 5.2725427660191364e-06, + "loss": 0.1339, + "step": 10160 + }, + { + "epoch": 1.4743575803704108, + "grad_norm": 1.71875, + "learning_rate": 5.258045810379821e-06, + "loss": 0.1327, + "step": 10170 + }, + { + "epoch": 1.475807328476677, + "grad_norm": 1.4375, + "learning_rate": 5.243548854740504e-06, + "loss": 0.1456, + "step": 10180 + }, + { + "epoch": 1.4772570765829438, + "grad_norm": 1.3125, + "learning_rate": 5.22905189910119e-06, + "loss": 0.1408, + "step": 10190 + }, + { + "epoch": 1.4787068246892103, + "grad_norm": 1.4296875, + "learning_rate": 5.214554943461873e-06, + "loss": 0.1486, + "step": 10200 + }, + { + "epoch": 1.4801565727954769, + "grad_norm": 2.109375, + "learning_rate": 5.200057987822558e-06, + "loss": 0.1152, + "step": 10210 + }, + { + "epoch": 1.4816063209017434, + "grad_norm": 2.015625, + "learning_rate": 5.185561032183242e-06, + "loss": 0.1304, + "step": 10220 + }, + { + "epoch": 1.48305606900801, + "grad_norm": 1.578125, + "learning_rate": 5.171064076543926e-06, + "loss": 0.1556, + "step": 10230 + }, + { + "epoch": 1.4845058171142764, + "grad_norm": 1.84375, + "learning_rate": 5.156567120904611e-06, + "loss": 0.167, + "step": 10240 + }, + { + "epoch": 1.485955565220543, + "grad_norm": 1.6953125, + "learning_rate": 5.1420701652652945e-06, + "loss": 0.1518, + "step": 10250 + }, + { + "epoch": 1.4874053133268095, + "grad_norm": 1.5390625, + "learning_rate": 5.127573209625979e-06, + "loss": 0.1354, + "step": 10260 + }, + { + "epoch": 1.488855061433076, + "grad_norm": 1.5, + "learning_rate": 5.113076253986663e-06, + "loss": 0.1252, + "step": 10270 + }, + { + "epoch": 1.4903048095393425, + "grad_norm": 1.3828125, + "learning_rate": 5.098579298347347e-06, + "loss": 0.1391, + "step": 10280 + }, + { + "epoch": 1.491754557645609, + "grad_norm": 1.15625, + "learning_rate": 5.084082342708032e-06, + "loss": 0.1138, + "step": 10290 + }, + { + "epoch": 1.4932043057518756, + "grad_norm": 1.8828125, + "learning_rate": 5.0695853870687166e-06, + "loss": 0.1612, + "step": 10300 + }, + { + "epoch": 1.494654053858142, + "grad_norm": 2.078125, + "learning_rate": 5.0550884314294e-06, + "loss": 0.141, + "step": 10310 + }, + { + "epoch": 1.4961038019644086, + "grad_norm": 1.8515625, + "learning_rate": 5.0405914757900846e-06, + "loss": 0.1422, + "step": 10320 + }, + { + "epoch": 1.4975535500706751, + "grad_norm": 1.7578125, + "learning_rate": 5.026094520150768e-06, + "loss": 0.1298, + "step": 10330 + }, + { + "epoch": 1.499003298176942, + "grad_norm": 1.6796875, + "learning_rate": 5.011597564511453e-06, + "loss": 0.1318, + "step": 10340 + }, + { + "epoch": 1.5004530462832082, + "grad_norm": 1.796875, + "learning_rate": 4.997100608872138e-06, + "loss": 0.1643, + "step": 10350 + }, + { + "epoch": 1.501902794389475, + "grad_norm": 1.578125, + "learning_rate": 4.982603653232821e-06, + "loss": 0.1367, + "step": 10360 + }, + { + "epoch": 1.5033525424957412, + "grad_norm": 1.6015625, + "learning_rate": 4.968106697593506e-06, + "loss": 0.1388, + "step": 10370 + }, + { + "epoch": 1.504802290602008, + "grad_norm": 1.6171875, + "learning_rate": 4.95360974195419e-06, + "loss": 0.1498, + "step": 10380 + }, + { + "epoch": 1.5062520387082743, + "grad_norm": 1.875, + "learning_rate": 4.939112786314874e-06, + "loss": 0.1285, + "step": 10390 + }, + { + "epoch": 1.507701786814541, + "grad_norm": 1.28125, + "learning_rate": 4.924615830675559e-06, + "loss": 0.1401, + "step": 10400 + }, + { + "epoch": 1.5091515349208076, + "grad_norm": 1.65625, + "learning_rate": 4.910118875036243e-06, + "loss": 0.1459, + "step": 10410 + }, + { + "epoch": 1.510601283027074, + "grad_norm": 2.03125, + "learning_rate": 4.895621919396927e-06, + "loss": 0.1283, + "step": 10420 + }, + { + "epoch": 1.5120510311333406, + "grad_norm": 1.25, + "learning_rate": 4.8811249637576114e-06, + "loss": 0.1435, + "step": 10430 + }, + { + "epoch": 1.5135007792396071, + "grad_norm": 1.4921875, + "learning_rate": 4.866628008118295e-06, + "loss": 0.1174, + "step": 10440 + }, + { + "epoch": 1.5149505273458737, + "grad_norm": 1.53125, + "learning_rate": 4.85213105247898e-06, + "loss": 0.1323, + "step": 10450 + }, + { + "epoch": 1.5164002754521402, + "grad_norm": 1.6796875, + "learning_rate": 4.837634096839664e-06, + "loss": 0.1413, + "step": 10460 + }, + { + "epoch": 1.5178500235584067, + "grad_norm": 1.03125, + "learning_rate": 4.823137141200348e-06, + "loss": 0.1294, + "step": 10470 + }, + { + "epoch": 1.5192997716646732, + "grad_norm": 0.921875, + "learning_rate": 4.808640185561033e-06, + "loss": 0.1528, + "step": 10480 + }, + { + "epoch": 1.5207495197709398, + "grad_norm": 1.3671875, + "learning_rate": 4.794143229921716e-06, + "loss": 0.1331, + "step": 10490 + }, + { + "epoch": 1.5221992678772063, + "grad_norm": 1.4921875, + "learning_rate": 4.7796462742824015e-06, + "loss": 0.1311, + "step": 10500 + }, + { + "epoch": 1.523649015983473, + "grad_norm": 2.296875, + "learning_rate": 4.765149318643085e-06, + "loss": 0.1483, + "step": 10510 + }, + { + "epoch": 1.5250987640897393, + "grad_norm": 1.703125, + "learning_rate": 4.7506523630037695e-06, + "loss": 0.1527, + "step": 10520 + }, + { + "epoch": 1.526548512196006, + "grad_norm": 1.5703125, + "learning_rate": 4.736155407364454e-06, + "loss": 0.1228, + "step": 10530 + }, + { + "epoch": 1.5279982603022724, + "grad_norm": 1.890625, + "learning_rate": 4.721658451725138e-06, + "loss": 0.114, + "step": 10540 + }, + { + "epoch": 1.5294480084085391, + "grad_norm": 1.5078125, + "learning_rate": 4.707161496085823e-06, + "loss": 0.1416, + "step": 10550 + }, + { + "epoch": 1.5308977565148054, + "grad_norm": 1.5390625, + "learning_rate": 4.692664540446506e-06, + "loss": 0.1529, + "step": 10560 + }, + { + "epoch": 1.5323475046210722, + "grad_norm": 1.890625, + "learning_rate": 4.678167584807191e-06, + "loss": 0.1587, + "step": 10570 + }, + { + "epoch": 1.5337972527273385, + "grad_norm": 1.6328125, + "learning_rate": 4.663670629167875e-06, + "loss": 0.137, + "step": 10580 + }, + { + "epoch": 1.5352470008336052, + "grad_norm": 1.265625, + "learning_rate": 4.6491736735285596e-06, + "loss": 0.1449, + "step": 10590 + }, + { + "epoch": 1.5366967489398717, + "grad_norm": 1.8203125, + "learning_rate": 4.634676717889244e-06, + "loss": 0.122, + "step": 10600 + }, + { + "epoch": 1.5381464970461383, + "grad_norm": 1.8046875, + "learning_rate": 4.6201797622499276e-06, + "loss": 0.142, + "step": 10610 + }, + { + "epoch": 1.5395962451524048, + "grad_norm": 2.03125, + "learning_rate": 4.605682806610613e-06, + "loss": 0.1328, + "step": 10620 + }, + { + "epoch": 1.5410459932586713, + "grad_norm": 1.1796875, + "learning_rate": 4.591185850971296e-06, + "loss": 0.1382, + "step": 10630 + }, + { + "epoch": 1.5424957413649378, + "grad_norm": 1.7890625, + "learning_rate": 4.576688895331981e-06, + "loss": 0.1538, + "step": 10640 + }, + { + "epoch": 1.5439454894712044, + "grad_norm": 2.0625, + "learning_rate": 4.562191939692665e-06, + "loss": 0.1488, + "step": 10650 + }, + { + "epoch": 1.5453952375774709, + "grad_norm": 1.9765625, + "learning_rate": 4.547694984053349e-06, + "loss": 0.1379, + "step": 10660 + }, + { + "epoch": 1.5468449856837374, + "grad_norm": 1.8671875, + "learning_rate": 4.533198028414034e-06, + "loss": 0.1471, + "step": 10670 + }, + { + "epoch": 1.5482947337900042, + "grad_norm": 1.0078125, + "learning_rate": 4.518701072774718e-06, + "loss": 0.1446, + "step": 10680 + }, + { + "epoch": 1.5497444818962705, + "grad_norm": 1.6640625, + "learning_rate": 4.504204117135402e-06, + "loss": 0.1277, + "step": 10690 + }, + { + "epoch": 1.5511942300025372, + "grad_norm": 1.203125, + "learning_rate": 4.4897071614960865e-06, + "loss": 0.1534, + "step": 10700 + }, + { + "epoch": 1.5526439781088035, + "grad_norm": 1.515625, + "learning_rate": 4.47521020585677e-06, + "loss": 0.1328, + "step": 10710 + }, + { + "epoch": 1.5540937262150702, + "grad_norm": 2.234375, + "learning_rate": 4.460713250217455e-06, + "loss": 0.1544, + "step": 10720 + }, + { + "epoch": 1.5555434743213366, + "grad_norm": 1.671875, + "learning_rate": 4.446216294578139e-06, + "loss": 0.1258, + "step": 10730 + }, + { + "epoch": 1.5569932224276033, + "grad_norm": 1.890625, + "learning_rate": 4.431719338938823e-06, + "loss": 0.126, + "step": 10740 + }, + { + "epoch": 1.5584429705338696, + "grad_norm": 1.453125, + "learning_rate": 4.417222383299508e-06, + "loss": 0.1155, + "step": 10750 + }, + { + "epoch": 1.5598927186401363, + "grad_norm": 1.5, + "learning_rate": 4.402725427660191e-06, + "loss": 0.1544, + "step": 10760 + }, + { + "epoch": 1.5613424667464026, + "grad_norm": 2.609375, + "learning_rate": 4.3882284720208765e-06, + "loss": 0.1532, + "step": 10770 + }, + { + "epoch": 1.5627922148526694, + "grad_norm": 1.5, + "learning_rate": 4.37373151638156e-06, + "loss": 0.1228, + "step": 10780 + }, + { + "epoch": 1.564241962958936, + "grad_norm": 2.140625, + "learning_rate": 4.3592345607422445e-06, + "loss": 0.1487, + "step": 10790 + }, + { + "epoch": 1.5656917110652024, + "grad_norm": 1.4375, + "learning_rate": 4.344737605102929e-06, + "loss": 0.1615, + "step": 10800 + }, + { + "epoch": 1.567141459171469, + "grad_norm": 1.2890625, + "learning_rate": 4.3302406494636125e-06, + "loss": 0.1294, + "step": 10810 + }, + { + "epoch": 1.5685912072777355, + "grad_norm": 1.0703125, + "learning_rate": 4.315743693824298e-06, + "loss": 0.1238, + "step": 10820 + }, + { + "epoch": 1.570040955384002, + "grad_norm": 1.515625, + "learning_rate": 4.301246738184981e-06, + "loss": 0.1444, + "step": 10830 + }, + { + "epoch": 1.5714907034902685, + "grad_norm": 1.140625, + "learning_rate": 4.286749782545666e-06, + "loss": 0.1346, + "step": 10840 + }, + { + "epoch": 1.572940451596535, + "grad_norm": 1.6875, + "learning_rate": 4.27225282690635e-06, + "loss": 0.1442, + "step": 10850 + }, + { + "epoch": 1.5743901997028016, + "grad_norm": 1.6796875, + "learning_rate": 4.257755871267034e-06, + "loss": 0.1232, + "step": 10860 + }, + { + "epoch": 1.5758399478090683, + "grad_norm": 1.3125, + "learning_rate": 4.243258915627719e-06, + "loss": 0.132, + "step": 10870 + }, + { + "epoch": 1.5772896959153346, + "grad_norm": 1.5, + "learning_rate": 4.2287619599884026e-06, + "loss": 0.13, + "step": 10880 + }, + { + "epoch": 1.5787394440216014, + "grad_norm": 1.5078125, + "learning_rate": 4.214265004349087e-06, + "loss": 0.1422, + "step": 10890 + }, + { + "epoch": 1.5801891921278677, + "grad_norm": 2.09375, + "learning_rate": 4.199768048709771e-06, + "loss": 0.1499, + "step": 10900 + }, + { + "epoch": 1.5816389402341344, + "grad_norm": 2.078125, + "learning_rate": 4.185271093070456e-06, + "loss": 0.1508, + "step": 10910 + }, + { + "epoch": 1.5830886883404007, + "grad_norm": 1.875, + "learning_rate": 4.17077413743114e-06, + "loss": 0.1362, + "step": 10920 + }, + { + "epoch": 1.5845384364466675, + "grad_norm": 1.6640625, + "learning_rate": 4.156277181791824e-06, + "loss": 0.1366, + "step": 10930 + }, + { + "epoch": 1.5859881845529338, + "grad_norm": 1.984375, + "learning_rate": 4.141780226152508e-06, + "loss": 0.1423, + "step": 10940 + }, + { + "epoch": 1.5874379326592005, + "grad_norm": 1.8046875, + "learning_rate": 4.127283270513193e-06, + "loss": 0.1236, + "step": 10950 + }, + { + "epoch": 1.588887680765467, + "grad_norm": 1.5390625, + "learning_rate": 4.112786314873877e-06, + "loss": 0.1224, + "step": 10960 + }, + { + "epoch": 1.5903374288717336, + "grad_norm": 1.765625, + "learning_rate": 4.0982893592345615e-06, + "loss": 0.1482, + "step": 10970 + }, + { + "epoch": 1.591787176978, + "grad_norm": 1.8984375, + "learning_rate": 4.083792403595245e-06, + "loss": 0.1358, + "step": 10980 + }, + { + "epoch": 1.5932369250842666, + "grad_norm": 1.640625, + "learning_rate": 4.0692954479559295e-06, + "loss": 0.1401, + "step": 10990 + }, + { + "epoch": 1.5946866731905331, + "grad_norm": 1.828125, + "learning_rate": 4.054798492316614e-06, + "loss": 0.1464, + "step": 11000 + }, + { + "epoch": 1.5961364212967997, + "grad_norm": 1.828125, + "learning_rate": 4.040301536677298e-06, + "loss": 0.1673, + "step": 11010 + }, + { + "epoch": 1.5975861694030662, + "grad_norm": 1.1484375, + "learning_rate": 4.025804581037983e-06, + "loss": 0.1207, + "step": 11020 + }, + { + "epoch": 1.5990359175093327, + "grad_norm": 1.84375, + "learning_rate": 4.011307625398666e-06, + "loss": 0.1557, + "step": 11030 + }, + { + "epoch": 1.6004856656155992, + "grad_norm": 1.6328125, + "learning_rate": 3.996810669759351e-06, + "loss": 0.1062, + "step": 11040 + }, + { + "epoch": 1.6019354137218658, + "grad_norm": 1.4609375, + "learning_rate": 3.982313714120035e-06, + "loss": 0.1381, + "step": 11050 + }, + { + "epoch": 1.6033851618281325, + "grad_norm": 1.5234375, + "learning_rate": 3.9678167584807195e-06, + "loss": 0.1251, + "step": 11060 + }, + { + "epoch": 1.6048349099343988, + "grad_norm": 1.40625, + "learning_rate": 3.953319802841404e-06, + "loss": 0.1105, + "step": 11070 + }, + { + "epoch": 1.6062846580406656, + "grad_norm": 1.8125, + "learning_rate": 3.9388228472020875e-06, + "loss": 0.1526, + "step": 11080 + }, + { + "epoch": 1.6077344061469319, + "grad_norm": 1.625, + "learning_rate": 3.924325891562772e-06, + "loss": 0.1023, + "step": 11090 + }, + { + "epoch": 1.6091841542531986, + "grad_norm": 1.375, + "learning_rate": 3.909828935923456e-06, + "loss": 0.1256, + "step": 11100 + }, + { + "epoch": 1.610633902359465, + "grad_norm": 1.6171875, + "learning_rate": 3.895331980284141e-06, + "loss": 0.1316, + "step": 11110 + }, + { + "epoch": 1.6120836504657317, + "grad_norm": 1.5859375, + "learning_rate": 3.880835024644825e-06, + "loss": 0.1149, + "step": 11120 + }, + { + "epoch": 1.613533398571998, + "grad_norm": 1.9296875, + "learning_rate": 3.866338069005509e-06, + "loss": 0.1566, + "step": 11130 + }, + { + "epoch": 1.6149831466782647, + "grad_norm": 1.7421875, + "learning_rate": 3.851841113366193e-06, + "loss": 0.1409, + "step": 11140 + }, + { + "epoch": 1.6164328947845312, + "grad_norm": 1.984375, + "learning_rate": 3.837344157726878e-06, + "loss": 0.1386, + "step": 11150 + }, + { + "epoch": 1.6178826428907978, + "grad_norm": 1.78125, + "learning_rate": 3.822847202087562e-06, + "loss": 0.1384, + "step": 11160 + }, + { + "epoch": 1.6193323909970643, + "grad_norm": 1.0859375, + "learning_rate": 3.8083502464482464e-06, + "loss": 0.1484, + "step": 11170 + }, + { + "epoch": 1.6207821391033308, + "grad_norm": 1.6953125, + "learning_rate": 3.7938532908089304e-06, + "loss": 0.1436, + "step": 11180 + }, + { + "epoch": 1.6222318872095973, + "grad_norm": 1.703125, + "learning_rate": 3.7793563351696144e-06, + "loss": 0.1075, + "step": 11190 + }, + { + "epoch": 1.6236816353158638, + "grad_norm": 1.5625, + "learning_rate": 3.764859379530299e-06, + "loss": 0.1419, + "step": 11200 + }, + { + "epoch": 1.6251313834221304, + "grad_norm": 1.1875, + "learning_rate": 3.7503624238909832e-06, + "loss": 0.1185, + "step": 11210 + }, + { + "epoch": 1.626581131528397, + "grad_norm": 2.171875, + "learning_rate": 3.7358654682516677e-06, + "loss": 0.1326, + "step": 11220 + }, + { + "epoch": 1.6280308796346636, + "grad_norm": 1.609375, + "learning_rate": 3.7213685126123516e-06, + "loss": 0.1144, + "step": 11230 + }, + { + "epoch": 1.62948062774093, + "grad_norm": 2.234375, + "learning_rate": 3.7068715569730356e-06, + "loss": 0.161, + "step": 11240 + }, + { + "epoch": 1.6309303758471967, + "grad_norm": 1.828125, + "learning_rate": 3.6923746013337205e-06, + "loss": 0.1161, + "step": 11250 + }, + { + "epoch": 1.632380123953463, + "grad_norm": 1.7265625, + "learning_rate": 3.6778776456944045e-06, + "loss": 0.1494, + "step": 11260 + }, + { + "epoch": 1.6338298720597297, + "grad_norm": 1.359375, + "learning_rate": 3.663380690055089e-06, + "loss": 0.1405, + "step": 11270 + }, + { + "epoch": 1.635279620165996, + "grad_norm": 1.734375, + "learning_rate": 3.648883734415773e-06, + "loss": 0.12, + "step": 11280 + }, + { + "epoch": 1.6367293682722628, + "grad_norm": 1.6328125, + "learning_rate": 3.6343867787764577e-06, + "loss": 0.1451, + "step": 11290 + }, + { + "epoch": 1.638179116378529, + "grad_norm": 1.8671875, + "learning_rate": 3.6198898231371417e-06, + "loss": 0.1135, + "step": 11300 + }, + { + "epoch": 1.6396288644847958, + "grad_norm": 1.6953125, + "learning_rate": 3.6053928674978257e-06, + "loss": 0.1153, + "step": 11310 + }, + { + "epoch": 1.6410786125910624, + "grad_norm": 0.79296875, + "learning_rate": 3.59089591185851e-06, + "loss": 0.1381, + "step": 11320 + }, + { + "epoch": 1.6425283606973289, + "grad_norm": 1.5078125, + "learning_rate": 3.576398956219194e-06, + "loss": 0.1327, + "step": 11330 + }, + { + "epoch": 1.6439781088035954, + "grad_norm": 1.0546875, + "learning_rate": 3.561902000579879e-06, + "loss": 0.137, + "step": 11340 + }, + { + "epoch": 1.645427856909862, + "grad_norm": 1.8359375, + "learning_rate": 3.547405044940563e-06, + "loss": 0.1286, + "step": 11350 + }, + { + "epoch": 1.6468776050161285, + "grad_norm": 2.15625, + "learning_rate": 3.532908089301247e-06, + "loss": 0.1207, + "step": 11360 + }, + { + "epoch": 1.648327353122395, + "grad_norm": 1.7421875, + "learning_rate": 3.5184111336619314e-06, + "loss": 0.1441, + "step": 11370 + }, + { + "epoch": 1.6497771012286615, + "grad_norm": 1.6328125, + "learning_rate": 3.5039141780226154e-06, + "loss": 0.116, + "step": 11380 + }, + { + "epoch": 1.651226849334928, + "grad_norm": 2.15625, + "learning_rate": 3.4894172223833e-06, + "loss": 0.1413, + "step": 11390 + }, + { + "epoch": 1.6526765974411946, + "grad_norm": 1.6328125, + "learning_rate": 3.474920266743984e-06, + "loss": 0.1392, + "step": 11400 + }, + { + "epoch": 1.654126345547461, + "grad_norm": 1.5390625, + "learning_rate": 3.460423311104668e-06, + "loss": 0.1461, + "step": 11410 + }, + { + "epoch": 1.6555760936537278, + "grad_norm": 1.4921875, + "learning_rate": 3.4459263554653526e-06, + "loss": 0.1283, + "step": 11420 + }, + { + "epoch": 1.6570258417599941, + "grad_norm": 1.03125, + "learning_rate": 3.4314293998260366e-06, + "loss": 0.1343, + "step": 11430 + }, + { + "epoch": 1.6584755898662609, + "grad_norm": 1.640625, + "learning_rate": 3.4169324441867214e-06, + "loss": 0.1554, + "step": 11440 + }, + { + "epoch": 1.6599253379725272, + "grad_norm": 1.4296875, + "learning_rate": 3.4024354885474054e-06, + "loss": 0.1414, + "step": 11450 + }, + { + "epoch": 1.661375086078794, + "grad_norm": 1.6171875, + "learning_rate": 3.3879385329080894e-06, + "loss": 0.1368, + "step": 11460 + }, + { + "epoch": 1.6628248341850602, + "grad_norm": 2.625, + "learning_rate": 3.373441577268774e-06, + "loss": 0.1421, + "step": 11470 + }, + { + "epoch": 1.664274582291327, + "grad_norm": 1.953125, + "learning_rate": 3.358944621629458e-06, + "loss": 0.146, + "step": 11480 + }, + { + "epoch": 1.6657243303975933, + "grad_norm": 1.1484375, + "learning_rate": 3.3444476659901427e-06, + "loss": 0.1229, + "step": 11490 + }, + { + "epoch": 1.66717407850386, + "grad_norm": 1.3984375, + "learning_rate": 3.3299507103508267e-06, + "loss": 0.1294, + "step": 11500 + }, + { + "epoch": 1.6686238266101265, + "grad_norm": 1.2890625, + "learning_rate": 3.3154537547115107e-06, + "loss": 0.1153, + "step": 11510 + }, + { + "epoch": 1.670073574716393, + "grad_norm": 1.5703125, + "learning_rate": 3.300956799072195e-06, + "loss": 0.156, + "step": 11520 + }, + { + "epoch": 1.6715233228226596, + "grad_norm": 1.875, + "learning_rate": 3.286459843432879e-06, + "loss": 0.1423, + "step": 11530 + }, + { + "epoch": 1.672973070928926, + "grad_norm": 1.0234375, + "learning_rate": 3.271962887793564e-06, + "loss": 0.1286, + "step": 11540 + }, + { + "epoch": 1.6744228190351926, + "grad_norm": 1.375, + "learning_rate": 3.257465932154248e-06, + "loss": 0.1358, + "step": 11550 + }, + { + "epoch": 1.6758725671414592, + "grad_norm": 1.6484375, + "learning_rate": 3.242968976514932e-06, + "loss": 0.1352, + "step": 11560 + }, + { + "epoch": 1.6773223152477257, + "grad_norm": 1.421875, + "learning_rate": 3.2284720208756163e-06, + "loss": 0.154, + "step": 11570 + }, + { + "epoch": 1.6787720633539922, + "grad_norm": 1.7734375, + "learning_rate": 3.2139750652363007e-06, + "loss": 0.1541, + "step": 11580 + }, + { + "epoch": 1.6802218114602587, + "grad_norm": 1.265625, + "learning_rate": 3.199478109596985e-06, + "loss": 0.1346, + "step": 11590 + }, + { + "epoch": 1.6816715595665253, + "grad_norm": 1.8671875, + "learning_rate": 3.184981153957669e-06, + "loss": 0.1175, + "step": 11600 + }, + { + "epoch": 1.683121307672792, + "grad_norm": 1.8984375, + "learning_rate": 3.170484198318353e-06, + "loss": 0.1327, + "step": 11610 + }, + { + "epoch": 1.6845710557790583, + "grad_norm": 1.921875, + "learning_rate": 3.155987242679038e-06, + "loss": 0.1256, + "step": 11620 + }, + { + "epoch": 1.686020803885325, + "grad_norm": 2.0625, + "learning_rate": 3.141490287039722e-06, + "loss": 0.1291, + "step": 11630 + }, + { + "epoch": 1.6874705519915913, + "grad_norm": 2.15625, + "learning_rate": 3.1269933314004064e-06, + "loss": 0.1423, + "step": 11640 + }, + { + "epoch": 1.688920300097858, + "grad_norm": 1.8046875, + "learning_rate": 3.1124963757610904e-06, + "loss": 0.1358, + "step": 11650 + }, + { + "epoch": 1.6903700482041244, + "grad_norm": 0.87109375, + "learning_rate": 3.0979994201217744e-06, + "loss": 0.1407, + "step": 11660 + }, + { + "epoch": 1.6918197963103911, + "grad_norm": 2.34375, + "learning_rate": 3.083502464482459e-06, + "loss": 0.1258, + "step": 11670 + }, + { + "epoch": 1.6932695444166574, + "grad_norm": 0.98828125, + "learning_rate": 3.069005508843143e-06, + "loss": 0.1235, + "step": 11680 + }, + { + "epoch": 1.6947192925229242, + "grad_norm": 1.59375, + "learning_rate": 3.0545085532038276e-06, + "loss": 0.1391, + "step": 11690 + }, + { + "epoch": 1.6961690406291907, + "grad_norm": 1.0390625, + "learning_rate": 3.0400115975645116e-06, + "loss": 0.1025, + "step": 11700 + }, + { + "epoch": 1.6976187887354572, + "grad_norm": 1.59375, + "learning_rate": 3.0255146419251956e-06, + "loss": 0.1496, + "step": 11710 + }, + { + "epoch": 1.6990685368417238, + "grad_norm": 1.65625, + "learning_rate": 3.0110176862858804e-06, + "loss": 0.1331, + "step": 11720 + }, + { + "epoch": 1.7005182849479903, + "grad_norm": 1.96875, + "learning_rate": 2.9965207306465644e-06, + "loss": 0.1492, + "step": 11730 + }, + { + "epoch": 1.7019680330542568, + "grad_norm": 1.421875, + "learning_rate": 2.982023775007249e-06, + "loss": 0.1269, + "step": 11740 + }, + { + "epoch": 1.7034177811605233, + "grad_norm": 1.5234375, + "learning_rate": 2.967526819367933e-06, + "loss": 0.139, + "step": 11750 + }, + { + "epoch": 1.7048675292667899, + "grad_norm": 1.6328125, + "learning_rate": 2.953029863728617e-06, + "loss": 0.1191, + "step": 11760 + }, + { + "epoch": 1.7063172773730564, + "grad_norm": 1.84375, + "learning_rate": 2.9385329080893017e-06, + "loss": 0.1379, + "step": 11770 + }, + { + "epoch": 1.7077670254793231, + "grad_norm": 1.875, + "learning_rate": 2.9240359524499857e-06, + "loss": 0.1262, + "step": 11780 + }, + { + "epoch": 1.7092167735855894, + "grad_norm": 1.6953125, + "learning_rate": 2.90953899681067e-06, + "loss": 0.1282, + "step": 11790 + }, + { + "epoch": 1.7106665216918562, + "grad_norm": 1.671875, + "learning_rate": 2.895042041171354e-06, + "loss": 0.1232, + "step": 11800 + }, + { + "epoch": 1.7121162697981225, + "grad_norm": 1.6640625, + "learning_rate": 2.880545085532038e-06, + "loss": 0.1099, + "step": 11810 + }, + { + "epoch": 1.7135660179043892, + "grad_norm": 1.1796875, + "learning_rate": 2.866048129892723e-06, + "loss": 0.1261, + "step": 11820 + }, + { + "epoch": 1.7150157660106555, + "grad_norm": 1.2421875, + "learning_rate": 2.851551174253407e-06, + "loss": 0.1532, + "step": 11830 + }, + { + "epoch": 1.7164655141169223, + "grad_norm": 1.8515625, + "learning_rate": 2.8370542186140913e-06, + "loss": 0.1223, + "step": 11840 + }, + { + "epoch": 1.7179152622231886, + "grad_norm": 1.484375, + "learning_rate": 2.8225572629747753e-06, + "loss": 0.1288, + "step": 11850 + }, + { + "epoch": 1.7193650103294553, + "grad_norm": 2.265625, + "learning_rate": 2.8080603073354597e-06, + "loss": 0.1428, + "step": 11860 + }, + { + "epoch": 1.7208147584357218, + "grad_norm": 1.59375, + "learning_rate": 2.793563351696144e-06, + "loss": 0.1154, + "step": 11870 + }, + { + "epoch": 1.7222645065419884, + "grad_norm": 1.46875, + "learning_rate": 2.779066396056828e-06, + "loss": 0.1417, + "step": 11880 + }, + { + "epoch": 1.723714254648255, + "grad_norm": 1.859375, + "learning_rate": 2.7645694404175126e-06, + "loss": 0.1342, + "step": 11890 + }, + { + "epoch": 1.7251640027545214, + "grad_norm": 2.203125, + "learning_rate": 2.750072484778197e-06, + "loss": 0.1345, + "step": 11900 + }, + { + "epoch": 1.726613750860788, + "grad_norm": 1.65625, + "learning_rate": 2.735575529138881e-06, + "loss": 0.1484, + "step": 11910 + }, + { + "epoch": 1.7280634989670545, + "grad_norm": 2.125, + "learning_rate": 2.7210785734995654e-06, + "loss": 0.1257, + "step": 11920 + }, + { + "epoch": 1.729513247073321, + "grad_norm": 1.4375, + "learning_rate": 2.7065816178602494e-06, + "loss": 0.1381, + "step": 11930 + }, + { + "epoch": 1.7309629951795875, + "grad_norm": 1.5546875, + "learning_rate": 2.692084662220934e-06, + "loss": 0.1379, + "step": 11940 + }, + { + "epoch": 1.732412743285854, + "grad_norm": 1.546875, + "learning_rate": 2.6775877065816182e-06, + "loss": 0.1357, + "step": 11950 + }, + { + "epoch": 1.7338624913921206, + "grad_norm": 2.328125, + "learning_rate": 2.663090750942302e-06, + "loss": 0.1259, + "step": 11960 + }, + { + "epoch": 1.7353122394983873, + "grad_norm": 1.7265625, + "learning_rate": 2.6485937953029866e-06, + "loss": 0.1379, + "step": 11970 + }, + { + "epoch": 1.7367619876046536, + "grad_norm": 1.6953125, + "learning_rate": 2.6340968396636706e-06, + "loss": 0.1237, + "step": 11980 + }, + { + "epoch": 1.7382117357109204, + "grad_norm": 1.4921875, + "learning_rate": 2.6195998840243555e-06, + "loss": 0.1126, + "step": 11990 + }, + { + "epoch": 1.7396614838171867, + "grad_norm": 1.609375, + "learning_rate": 2.6051029283850395e-06, + "loss": 0.1129, + "step": 12000 + }, + { + "epoch": 1.7411112319234534, + "grad_norm": 1.6953125, + "learning_rate": 2.5906059727457234e-06, + "loss": 0.1497, + "step": 12010 + }, + { + "epoch": 1.7425609800297197, + "grad_norm": 1.453125, + "learning_rate": 2.576109017106408e-06, + "loss": 0.114, + "step": 12020 + }, + { + "epoch": 1.7440107281359865, + "grad_norm": 1.0, + "learning_rate": 2.561612061467092e-06, + "loss": 0.136, + "step": 12030 + }, + { + "epoch": 1.7454604762422528, + "grad_norm": 1.7265625, + "learning_rate": 2.5471151058277767e-06, + "loss": 0.1292, + "step": 12040 + }, + { + "epoch": 1.7469102243485195, + "grad_norm": 1.8359375, + "learning_rate": 2.5326181501884607e-06, + "loss": 0.1402, + "step": 12050 + }, + { + "epoch": 1.748359972454786, + "grad_norm": 1.5703125, + "learning_rate": 2.5181211945491447e-06, + "loss": 0.1286, + "step": 12060 + }, + { + "epoch": 1.7498097205610526, + "grad_norm": 2.3125, + "learning_rate": 2.503624238909829e-06, + "loss": 0.1481, + "step": 12070 + }, + { + "epoch": 1.751259468667319, + "grad_norm": 1.8671875, + "learning_rate": 2.4891272832705135e-06, + "loss": 0.1437, + "step": 12080 + }, + { + "epoch": 1.7527092167735856, + "grad_norm": 1.921875, + "learning_rate": 2.4746303276311975e-06, + "loss": 0.1494, + "step": 12090 + }, + { + "epoch": 1.7541589648798521, + "grad_norm": 1.6953125, + "learning_rate": 2.460133371991882e-06, + "loss": 0.1301, + "step": 12100 + }, + { + "epoch": 1.7556087129861186, + "grad_norm": 1.53125, + "learning_rate": 2.4456364163525663e-06, + "loss": 0.1361, + "step": 12110 + }, + { + "epoch": 1.7570584610923852, + "grad_norm": 1.6328125, + "learning_rate": 2.4311394607132503e-06, + "loss": 0.143, + "step": 12120 + }, + { + "epoch": 1.7585082091986517, + "grad_norm": 1.3203125, + "learning_rate": 2.4166425050739348e-06, + "loss": 0.1049, + "step": 12130 + }, + { + "epoch": 1.7599579573049182, + "grad_norm": 1.7265625, + "learning_rate": 2.4021455494346187e-06, + "loss": 0.1151, + "step": 12140 + }, + { + "epoch": 1.7614077054111847, + "grad_norm": 1.2578125, + "learning_rate": 2.387648593795303e-06, + "loss": 0.1508, + "step": 12150 + }, + { + "epoch": 1.7628574535174515, + "grad_norm": 1.28125, + "learning_rate": 2.3731516381559876e-06, + "loss": 0.1403, + "step": 12160 + }, + { + "epoch": 1.7643072016237178, + "grad_norm": 1.5625, + "learning_rate": 2.3586546825166716e-06, + "loss": 0.1281, + "step": 12170 + }, + { + "epoch": 1.7657569497299845, + "grad_norm": 1.296875, + "learning_rate": 2.344157726877356e-06, + "loss": 0.1475, + "step": 12180 + }, + { + "epoch": 1.7672066978362508, + "grad_norm": 1.625, + "learning_rate": 2.32966077123804e-06, + "loss": 0.1149, + "step": 12190 + }, + { + "epoch": 1.7686564459425176, + "grad_norm": 1.609375, + "learning_rate": 2.3151638155987244e-06, + "loss": 0.1483, + "step": 12200 + }, + { + "epoch": 1.7701061940487839, + "grad_norm": 1.46875, + "learning_rate": 2.300666859959409e-06, + "loss": 0.1201, + "step": 12210 + }, + { + "epoch": 1.7715559421550506, + "grad_norm": 1.390625, + "learning_rate": 2.286169904320093e-06, + "loss": 0.1396, + "step": 12220 + }, + { + "epoch": 1.773005690261317, + "grad_norm": 1.9609375, + "learning_rate": 2.2716729486807772e-06, + "loss": 0.1519, + "step": 12230 + }, + { + "epoch": 1.7744554383675837, + "grad_norm": 1.953125, + "learning_rate": 2.2571759930414612e-06, + "loss": 0.1423, + "step": 12240 + }, + { + "epoch": 1.7759051864738502, + "grad_norm": 1.7421875, + "learning_rate": 2.2426790374021456e-06, + "loss": 0.113, + "step": 12250 + }, + { + "epoch": 1.7773549345801167, + "grad_norm": 1.4140625, + "learning_rate": 2.22818208176283e-06, + "loss": 0.1269, + "step": 12260 + }, + { + "epoch": 1.7788046826863833, + "grad_norm": 0.96484375, + "learning_rate": 2.2136851261235145e-06, + "loss": 0.1281, + "step": 12270 + }, + { + "epoch": 1.7802544307926498, + "grad_norm": 2.015625, + "learning_rate": 2.1991881704841985e-06, + "loss": 0.1453, + "step": 12280 + }, + { + "epoch": 1.7817041788989163, + "grad_norm": 2.03125, + "learning_rate": 2.184691214844883e-06, + "loss": 0.1505, + "step": 12290 + }, + { + "epoch": 1.7831539270051828, + "grad_norm": 2.0, + "learning_rate": 2.170194259205567e-06, + "loss": 0.1416, + "step": 12300 + }, + { + "epoch": 1.7846036751114494, + "grad_norm": 1.7109375, + "learning_rate": 2.1556973035662513e-06, + "loss": 0.1319, + "step": 12310 + }, + { + "epoch": 1.7860534232177159, + "grad_norm": 1.78125, + "learning_rate": 2.1412003479269357e-06, + "loss": 0.1489, + "step": 12320 + }, + { + "epoch": 1.7875031713239826, + "grad_norm": 1.984375, + "learning_rate": 2.1267033922876197e-06, + "loss": 0.1276, + "step": 12330 + }, + { + "epoch": 1.788952919430249, + "grad_norm": 2.359375, + "learning_rate": 2.112206436648304e-06, + "loss": 0.1364, + "step": 12340 + }, + { + "epoch": 1.7904026675365157, + "grad_norm": 1.328125, + "learning_rate": 2.097709481008988e-06, + "loss": 0.1295, + "step": 12350 + }, + { + "epoch": 1.791852415642782, + "grad_norm": 1.5859375, + "learning_rate": 2.0832125253696725e-06, + "loss": 0.1025, + "step": 12360 + }, + { + "epoch": 1.7933021637490487, + "grad_norm": 2.015625, + "learning_rate": 2.068715569730357e-06, + "loss": 0.1587, + "step": 12370 + }, + { + "epoch": 1.794751911855315, + "grad_norm": 2.5625, + "learning_rate": 2.054218614091041e-06, + "loss": 0.1266, + "step": 12380 + }, + { + "epoch": 1.7962016599615818, + "grad_norm": 1.65625, + "learning_rate": 2.0397216584517254e-06, + "loss": 0.1226, + "step": 12390 + }, + { + "epoch": 1.797651408067848, + "grad_norm": 1.4296875, + "learning_rate": 2.0252247028124093e-06, + "loss": 0.1379, + "step": 12400 + }, + { + "epoch": 1.7991011561741148, + "grad_norm": 1.546875, + "learning_rate": 2.0107277471730938e-06, + "loss": 0.1476, + "step": 12410 + }, + { + "epoch": 1.8005509042803813, + "grad_norm": 1.328125, + "learning_rate": 1.996230791533778e-06, + "loss": 0.1109, + "step": 12420 + }, + { + "epoch": 1.8020006523866479, + "grad_norm": 1.8984375, + "learning_rate": 1.9817338358944626e-06, + "loss": 0.1235, + "step": 12430 + }, + { + "epoch": 1.8034504004929144, + "grad_norm": 2.015625, + "learning_rate": 1.9672368802551466e-06, + "loss": 0.1436, + "step": 12440 + }, + { + "epoch": 1.804900148599181, + "grad_norm": 1.8984375, + "learning_rate": 1.9527399246158306e-06, + "loss": 0.1263, + "step": 12450 + }, + { + "epoch": 1.8063498967054474, + "grad_norm": 1.7265625, + "learning_rate": 1.938242968976515e-06, + "loss": 0.136, + "step": 12460 + }, + { + "epoch": 1.807799644811714, + "grad_norm": 1.8515625, + "learning_rate": 1.9237460133371994e-06, + "loss": 0.1145, + "step": 12470 + }, + { + "epoch": 1.8092493929179805, + "grad_norm": 2.0, + "learning_rate": 1.909249057697884e-06, + "loss": 0.1444, + "step": 12480 + }, + { + "epoch": 1.810699141024247, + "grad_norm": 1.8125, + "learning_rate": 1.894752102058568e-06, + "loss": 0.1392, + "step": 12490 + }, + { + "epoch": 1.8121488891305135, + "grad_norm": 1.8515625, + "learning_rate": 1.880255146419252e-06, + "loss": 0.1706, + "step": 12500 + }, + { + "epoch": 1.81359863723678, + "grad_norm": 1.359375, + "learning_rate": 1.8657581907799362e-06, + "loss": 0.1197, + "step": 12510 + }, + { + "epoch": 1.8150483853430468, + "grad_norm": 2.03125, + "learning_rate": 1.8512612351406207e-06, + "loss": 0.1442, + "step": 12520 + }, + { + "epoch": 1.816498133449313, + "grad_norm": 1.84375, + "learning_rate": 1.8367642795013049e-06, + "loss": 0.1238, + "step": 12530 + }, + { + "epoch": 1.8179478815555798, + "grad_norm": 1.609375, + "learning_rate": 1.8222673238619893e-06, + "loss": 0.1096, + "step": 12540 + }, + { + "epoch": 1.8193976296618461, + "grad_norm": 2.125, + "learning_rate": 1.8077703682226733e-06, + "loss": 0.1416, + "step": 12550 + }, + { + "epoch": 1.820847377768113, + "grad_norm": 1.375, + "learning_rate": 1.7932734125833575e-06, + "loss": 0.1347, + "step": 12560 + }, + { + "epoch": 1.8222971258743792, + "grad_norm": 2.390625, + "learning_rate": 1.7787764569440419e-06, + "loss": 0.1701, + "step": 12570 + }, + { + "epoch": 1.823746873980646, + "grad_norm": 1.828125, + "learning_rate": 1.764279501304726e-06, + "loss": 0.1391, + "step": 12580 + }, + { + "epoch": 1.8251966220869122, + "grad_norm": 1.2734375, + "learning_rate": 1.7497825456654105e-06, + "loss": 0.1238, + "step": 12590 + }, + { + "epoch": 1.826646370193179, + "grad_norm": 2.1875, + "learning_rate": 1.7352855900260945e-06, + "loss": 0.143, + "step": 12600 + }, + { + "epoch": 1.8280961182994455, + "grad_norm": 1.59375, + "learning_rate": 1.720788634386779e-06, + "loss": 0.1337, + "step": 12610 + }, + { + "epoch": 1.829545866405712, + "grad_norm": 1.84375, + "learning_rate": 1.7062916787474631e-06, + "loss": 0.1133, + "step": 12620 + }, + { + "epoch": 1.8309956145119786, + "grad_norm": 1.4609375, + "learning_rate": 1.6917947231081475e-06, + "loss": 0.1237, + "step": 12630 + }, + { + "epoch": 1.832445362618245, + "grad_norm": 2.09375, + "learning_rate": 1.6772977674688317e-06, + "loss": 0.1576, + "step": 12640 + }, + { + "epoch": 1.8338951107245116, + "grad_norm": 2.5625, + "learning_rate": 1.6628008118295157e-06, + "loss": 0.1394, + "step": 12650 + }, + { + "epoch": 1.8353448588307781, + "grad_norm": 1.65625, + "learning_rate": 1.6483038561902002e-06, + "loss": 0.1366, + "step": 12660 + }, + { + "epoch": 1.8367946069370447, + "grad_norm": 2.171875, + "learning_rate": 1.6338069005508844e-06, + "loss": 0.1576, + "step": 12670 + }, + { + "epoch": 1.8382443550433112, + "grad_norm": 2.296875, + "learning_rate": 1.6193099449115688e-06, + "loss": 0.1356, + "step": 12680 + }, + { + "epoch": 1.8396941031495777, + "grad_norm": 1.28125, + "learning_rate": 1.604812989272253e-06, + "loss": 0.135, + "step": 12690 + }, + { + "epoch": 1.8411438512558442, + "grad_norm": 1.390625, + "learning_rate": 1.5903160336329374e-06, + "loss": 0.1101, + "step": 12700 + }, + { + "epoch": 1.842593599362111, + "grad_norm": 1.5078125, + "learning_rate": 1.5758190779936214e-06, + "loss": 0.1127, + "step": 12710 + }, + { + "epoch": 1.8440433474683773, + "grad_norm": 1.3671875, + "learning_rate": 1.5613221223543056e-06, + "loss": 0.1302, + "step": 12720 + }, + { + "epoch": 1.845493095574644, + "grad_norm": 1.8984375, + "learning_rate": 1.54682516671499e-06, + "loss": 0.1287, + "step": 12730 + }, + { + "epoch": 1.8469428436809103, + "grad_norm": 1.5859375, + "learning_rate": 1.5323282110756742e-06, + "loss": 0.1195, + "step": 12740 + }, + { + "epoch": 1.848392591787177, + "grad_norm": 2.28125, + "learning_rate": 1.5178312554363586e-06, + "loss": 0.1354, + "step": 12750 + }, + { + "epoch": 1.8498423398934434, + "grad_norm": 1.9765625, + "learning_rate": 1.5033342997970426e-06, + "loss": 0.1462, + "step": 12760 + }, + { + "epoch": 1.8512920879997101, + "grad_norm": 2.6875, + "learning_rate": 1.488837344157727e-06, + "loss": 0.1494, + "step": 12770 + }, + { + "epoch": 1.8527418361059764, + "grad_norm": 1.1640625, + "learning_rate": 1.4743403885184113e-06, + "loss": 0.1612, + "step": 12780 + }, + { + "epoch": 1.8541915842122432, + "grad_norm": 1.46875, + "learning_rate": 1.4598434328790957e-06, + "loss": 0.1368, + "step": 12790 + }, + { + "epoch": 1.8556413323185097, + "grad_norm": 1.890625, + "learning_rate": 1.4453464772397799e-06, + "loss": 0.1053, + "step": 12800 + }, + { + "epoch": 1.8570910804247762, + "grad_norm": 1.9921875, + "learning_rate": 1.4308495216004639e-06, + "loss": 0.142, + "step": 12810 + }, + { + "epoch": 1.8585408285310427, + "grad_norm": 1.6171875, + "learning_rate": 1.4163525659611483e-06, + "loss": 0.1275, + "step": 12820 + }, + { + "epoch": 1.8599905766373093, + "grad_norm": 1.265625, + "learning_rate": 1.4018556103218325e-06, + "loss": 0.1309, + "step": 12830 + }, + { + "epoch": 1.8614403247435758, + "grad_norm": 0.7890625, + "learning_rate": 1.387358654682517e-06, + "loss": 0.1232, + "step": 12840 + }, + { + "epoch": 1.8628900728498423, + "grad_norm": 1.515625, + "learning_rate": 1.3728616990432011e-06, + "loss": 0.0871, + "step": 12850 + }, + { + "epoch": 1.8643398209561088, + "grad_norm": 1.796875, + "learning_rate": 1.358364743403885e-06, + "loss": 0.1547, + "step": 12860 + }, + { + "epoch": 1.8657895690623754, + "grad_norm": 1.7265625, + "learning_rate": 1.3438677877645695e-06, + "loss": 0.1127, + "step": 12870 + }, + { + "epoch": 1.867239317168642, + "grad_norm": 1.6640625, + "learning_rate": 1.3293708321252537e-06, + "loss": 0.1325, + "step": 12880 + }, + { + "epoch": 1.8686890652749084, + "grad_norm": 1.640625, + "learning_rate": 1.3148738764859381e-06, + "loss": 0.1103, + "step": 12890 + }, + { + "epoch": 1.8701388133811752, + "grad_norm": 1.859375, + "learning_rate": 1.3003769208466223e-06, + "loss": 0.1392, + "step": 12900 + }, + { + "epoch": 1.8715885614874415, + "grad_norm": 1.0546875, + "learning_rate": 1.2858799652073066e-06, + "loss": 0.1289, + "step": 12910 + }, + { + "epoch": 1.8730383095937082, + "grad_norm": 0.98828125, + "learning_rate": 1.2713830095679908e-06, + "loss": 0.1137, + "step": 12920 + }, + { + "epoch": 1.8744880576999745, + "grad_norm": 1.3984375, + "learning_rate": 1.2568860539286752e-06, + "loss": 0.1323, + "step": 12930 + }, + { + "epoch": 1.8759378058062413, + "grad_norm": 1.5859375, + "learning_rate": 1.2423890982893594e-06, + "loss": 0.1294, + "step": 12940 + }, + { + "epoch": 1.8773875539125076, + "grad_norm": 0.98828125, + "learning_rate": 1.2278921426500436e-06, + "loss": 0.1154, + "step": 12950 + }, + { + "epoch": 1.8788373020187743, + "grad_norm": 1.0625, + "learning_rate": 1.2133951870107278e-06, + "loss": 0.122, + "step": 12960 + }, + { + "epoch": 1.8802870501250408, + "grad_norm": 2.328125, + "learning_rate": 1.198898231371412e-06, + "loss": 0.1568, + "step": 12970 + }, + { + "epoch": 1.8817367982313074, + "grad_norm": 1.75, + "learning_rate": 1.1844012757320964e-06, + "loss": 0.1479, + "step": 12980 + }, + { + "epoch": 1.8831865463375739, + "grad_norm": 2.109375, + "learning_rate": 1.1699043200927806e-06, + "loss": 0.1343, + "step": 12990 + }, + { + "epoch": 1.8846362944438404, + "grad_norm": 1.5, + "learning_rate": 1.1554073644534648e-06, + "loss": 0.1237, + "step": 13000 + }, + { + "epoch": 1.886086042550107, + "grad_norm": 1.7265625, + "learning_rate": 1.1409104088141492e-06, + "loss": 0.1253, + "step": 13010 + }, + { + "epoch": 1.8875357906563734, + "grad_norm": 1.4375, + "learning_rate": 1.1264134531748334e-06, + "loss": 0.1332, + "step": 13020 + }, + { + "epoch": 1.88898553876264, + "grad_norm": 1.3203125, + "learning_rate": 1.1119164975355176e-06, + "loss": 0.1179, + "step": 13030 + }, + { + "epoch": 1.8904352868689065, + "grad_norm": 1.1328125, + "learning_rate": 1.0974195418962019e-06, + "loss": 0.134, + "step": 13040 + }, + { + "epoch": 1.891885034975173, + "grad_norm": 2.15625, + "learning_rate": 1.082922586256886e-06, + "loss": 0.1253, + "step": 13050 + }, + { + "epoch": 1.8933347830814395, + "grad_norm": 1.921875, + "learning_rate": 1.0684256306175705e-06, + "loss": 0.1246, + "step": 13060 + }, + { + "epoch": 1.8947845311877063, + "grad_norm": 1.8671875, + "learning_rate": 1.0539286749782547e-06, + "loss": 0.1266, + "step": 13070 + }, + { + "epoch": 1.8962342792939726, + "grad_norm": 1.421875, + "learning_rate": 1.0394317193389389e-06, + "loss": 0.1374, + "step": 13080 + }, + { + "epoch": 1.8976840274002393, + "grad_norm": 1.625, + "learning_rate": 1.024934763699623e-06, + "loss": 0.1556, + "step": 13090 + }, + { + "epoch": 1.8991337755065056, + "grad_norm": 1.8828125, + "learning_rate": 1.0104378080603073e-06, + "loss": 0.1232, + "step": 13100 + }, + { + "epoch": 1.9005835236127724, + "grad_norm": 1.390625, + "learning_rate": 9.959408524209917e-07, + "loss": 0.1189, + "step": 13110 + }, + { + "epoch": 1.9020332717190387, + "grad_norm": 2.109375, + "learning_rate": 9.81443896781676e-07, + "loss": 0.1498, + "step": 13120 + }, + { + "epoch": 1.9034830198253054, + "grad_norm": 1.2734375, + "learning_rate": 9.669469411423601e-07, + "loss": 0.1316, + "step": 13130 + }, + { + "epoch": 1.9049327679315717, + "grad_norm": 2.40625, + "learning_rate": 9.524499855030444e-07, + "loss": 0.1414, + "step": 13140 + }, + { + "epoch": 1.9063825160378385, + "grad_norm": 2.046875, + "learning_rate": 9.379530298637286e-07, + "loss": 0.1181, + "step": 13150 + }, + { + "epoch": 1.907832264144105, + "grad_norm": 1.25, + "learning_rate": 9.234560742244129e-07, + "loss": 0.1222, + "step": 13160 + }, + { + "epoch": 1.9092820122503715, + "grad_norm": 1.6875, + "learning_rate": 9.089591185850973e-07, + "loss": 0.1212, + "step": 13170 + }, + { + "epoch": 1.910731760356638, + "grad_norm": 1.234375, + "learning_rate": 8.944621629457815e-07, + "loss": 0.1413, + "step": 13180 + }, + { + "epoch": 1.9121815084629046, + "grad_norm": 1.8828125, + "learning_rate": 8.799652073064658e-07, + "loss": 0.1391, + "step": 13190 + }, + { + "epoch": 1.913631256569171, + "grad_norm": 1.4921875, + "learning_rate": 8.654682516671499e-07, + "loss": 0.1182, + "step": 13200 + }, + { + "epoch": 1.9150810046754376, + "grad_norm": 1.375, + "learning_rate": 8.509712960278342e-07, + "loss": 0.1083, + "step": 13210 + }, + { + "epoch": 1.9165307527817041, + "grad_norm": 1.765625, + "learning_rate": 8.364743403885185e-07, + "loss": 0.135, + "step": 13220 + }, + { + "epoch": 1.9179805008879707, + "grad_norm": 2.0, + "learning_rate": 8.219773847492027e-07, + "loss": 0.1493, + "step": 13230 + }, + { + "epoch": 1.9194302489942372, + "grad_norm": 2.0, + "learning_rate": 8.07480429109887e-07, + "loss": 0.1372, + "step": 13240 + }, + { + "epoch": 1.9208799971005037, + "grad_norm": 2.046875, + "learning_rate": 7.929834734705713e-07, + "loss": 0.1408, + "step": 13250 + }, + { + "epoch": 1.9223297452067705, + "grad_norm": 1.2421875, + "learning_rate": 7.784865178312555e-07, + "loss": 0.1256, + "step": 13260 + }, + { + "epoch": 1.9237794933130368, + "grad_norm": 1.8125, + "learning_rate": 7.639895621919398e-07, + "loss": 0.1378, + "step": 13270 + }, + { + "epoch": 1.9252292414193035, + "grad_norm": 1.90625, + "learning_rate": 7.494926065526239e-07, + "loss": 0.1188, + "step": 13280 + }, + { + "epoch": 1.9266789895255698, + "grad_norm": 2.03125, + "learning_rate": 7.349956509133082e-07, + "loss": 0.1508, + "step": 13290 + }, + { + "epoch": 1.9281287376318366, + "grad_norm": 1.7734375, + "learning_rate": 7.204986952739926e-07, + "loss": 0.1231, + "step": 13300 + }, + { + "epoch": 1.9295784857381029, + "grad_norm": 1.9375, + "learning_rate": 7.060017396346768e-07, + "loss": 0.1388, + "step": 13310 + }, + { + "epoch": 1.9310282338443696, + "grad_norm": 1.53125, + "learning_rate": 6.915047839953611e-07, + "loss": 0.1279, + "step": 13320 + }, + { + "epoch": 1.932477981950636, + "grad_norm": 1.375, + "learning_rate": 6.770078283560453e-07, + "loss": 0.139, + "step": 13330 + }, + { + "epoch": 1.9339277300569027, + "grad_norm": 1.828125, + "learning_rate": 6.625108727167296e-07, + "loss": 0.1505, + "step": 13340 + }, + { + "epoch": 1.9353774781631692, + "grad_norm": 1.78125, + "learning_rate": 6.480139170774138e-07, + "loss": 0.1573, + "step": 13350 + }, + { + "epoch": 1.9368272262694357, + "grad_norm": 1.7578125, + "learning_rate": 6.33516961438098e-07, + "loss": 0.1416, + "step": 13360 + }, + { + "epoch": 1.9382769743757022, + "grad_norm": 1.8515625, + "learning_rate": 6.190200057987823e-07, + "loss": 0.1391, + "step": 13370 + }, + { + "epoch": 1.9397267224819688, + "grad_norm": 1.515625, + "learning_rate": 6.045230501594665e-07, + "loss": 0.1382, + "step": 13380 + }, + { + "epoch": 1.9411764705882353, + "grad_norm": 2.09375, + "learning_rate": 5.900260945201508e-07, + "loss": 0.1424, + "step": 13390 + }, + { + "epoch": 1.9426262186945018, + "grad_norm": 1.28125, + "learning_rate": 5.75529138880835e-07, + "loss": 0.1134, + "step": 13400 + }, + { + "epoch": 1.9440759668007683, + "grad_norm": 0.9140625, + "learning_rate": 5.610321832415193e-07, + "loss": 0.124, + "step": 13410 + }, + { + "epoch": 1.9455257149070349, + "grad_norm": 2.0625, + "learning_rate": 5.465352276022035e-07, + "loss": 0.1528, + "step": 13420 + }, + { + "epoch": 1.9469754630133016, + "grad_norm": 1.53125, + "learning_rate": 5.320382719628879e-07, + "loss": 0.1109, + "step": 13430 + }, + { + "epoch": 1.948425211119568, + "grad_norm": 1.8203125, + "learning_rate": 5.175413163235721e-07, + "loss": 0.1667, + "step": 13440 + }, + { + "epoch": 1.9498749592258346, + "grad_norm": 1.984375, + "learning_rate": 5.030443606842563e-07, + "loss": 0.1352, + "step": 13450 + }, + { + "epoch": 1.951324707332101, + "grad_norm": 1.7578125, + "learning_rate": 4.885474050449406e-07, + "loss": 0.1284, + "step": 13460 + }, + { + "epoch": 1.9527744554383677, + "grad_norm": 1.2734375, + "learning_rate": 4.740504494056249e-07, + "loss": 0.1143, + "step": 13470 + }, + { + "epoch": 1.954224203544634, + "grad_norm": 1.90625, + "learning_rate": 4.595534937663091e-07, + "loss": 0.1445, + "step": 13480 + }, + { + "epoch": 1.9556739516509007, + "grad_norm": 1.4296875, + "learning_rate": 4.4505653812699335e-07, + "loss": 0.1381, + "step": 13490 + }, + { + "epoch": 1.957123699757167, + "grad_norm": 1.4921875, + "learning_rate": 4.3055958248767766e-07, + "loss": 0.1422, + "step": 13500 + }, + { + "epoch": 1.9585734478634338, + "grad_norm": 0.9453125, + "learning_rate": 4.1606262684836187e-07, + "loss": 0.1331, + "step": 13510 + }, + { + "epoch": 1.9600231959697003, + "grad_norm": 1.640625, + "learning_rate": 4.015656712090461e-07, + "loss": 0.1227, + "step": 13520 + }, + { + "epoch": 1.9614729440759668, + "grad_norm": 1.578125, + "learning_rate": 3.870687155697304e-07, + "loss": 0.1255, + "step": 13530 + }, + { + "epoch": 1.9629226921822334, + "grad_norm": 1.75, + "learning_rate": 3.725717599304146e-07, + "loss": 0.1109, + "step": 13540 + }, + { + "epoch": 1.9643724402884999, + "grad_norm": 1.9609375, + "learning_rate": 3.580748042910989e-07, + "loss": 0.1416, + "step": 13550 + }, + { + "epoch": 1.9658221883947664, + "grad_norm": 1.75, + "learning_rate": 3.4357784865178316e-07, + "loss": 0.1388, + "step": 13560 + }, + { + "epoch": 1.967271936501033, + "grad_norm": 1.8828125, + "learning_rate": 3.290808930124674e-07, + "loss": 0.1277, + "step": 13570 + }, + { + "epoch": 1.9687216846072995, + "grad_norm": 1.9609375, + "learning_rate": 3.145839373731516e-07, + "loss": 0.147, + "step": 13580 + }, + { + "epoch": 1.970171432713566, + "grad_norm": 1.6015625, + "learning_rate": 3.0008698173383593e-07, + "loss": 0.1309, + "step": 13590 + }, + { + "epoch": 1.9716211808198325, + "grad_norm": 1.6640625, + "learning_rate": 2.855900260945202e-07, + "loss": 0.1309, + "step": 13600 + }, + { + "epoch": 1.973070928926099, + "grad_norm": 1.6484375, + "learning_rate": 2.7109307045520445e-07, + "loss": 0.1506, + "step": 13610 + }, + { + "epoch": 1.9745206770323658, + "grad_norm": 1.8125, + "learning_rate": 2.5659611481588865e-07, + "loss": 0.1325, + "step": 13620 + }, + { + "epoch": 1.975970425138632, + "grad_norm": 1.6640625, + "learning_rate": 2.4209915917657296e-07, + "loss": 0.1172, + "step": 13630 + }, + { + "epoch": 1.9774201732448988, + "grad_norm": 2.09375, + "learning_rate": 2.276022035372572e-07, + "loss": 0.1442, + "step": 13640 + }, + { + "epoch": 1.9788699213511651, + "grad_norm": 1.75, + "learning_rate": 2.1310524789794145e-07, + "loss": 0.1307, + "step": 13650 + }, + { + "epoch": 1.9803196694574319, + "grad_norm": 1.6484375, + "learning_rate": 1.986082922586257e-07, + "loss": 0.1563, + "step": 13660 + }, + { + "epoch": 1.9817694175636982, + "grad_norm": 1.5234375, + "learning_rate": 1.8411133661930997e-07, + "loss": 0.1441, + "step": 13670 + }, + { + "epoch": 1.983219165669965, + "grad_norm": 1.40625, + "learning_rate": 1.6961438097999422e-07, + "loss": 0.1414, + "step": 13680 + }, + { + "epoch": 1.9846689137762312, + "grad_norm": 1.359375, + "learning_rate": 1.5511742534067848e-07, + "loss": 0.1397, + "step": 13690 + }, + { + "epoch": 1.986118661882498, + "grad_norm": 2.03125, + "learning_rate": 1.4062046970136271e-07, + "loss": 0.1318, + "step": 13700 + }, + { + "epoch": 1.9875684099887645, + "grad_norm": 2.109375, + "learning_rate": 1.2612351406204697e-07, + "loss": 0.1312, + "step": 13710 + }, + { + "epoch": 1.989018158095031, + "grad_norm": 1.2890625, + "learning_rate": 1.1162655842273124e-07, + "loss": 0.1324, + "step": 13720 + }, + { + "epoch": 1.9904679062012975, + "grad_norm": 1.359375, + "learning_rate": 9.71296027834155e-08, + "loss": 0.1352, + "step": 13730 + }, + { + "epoch": 1.991917654307564, + "grad_norm": 2.046875, + "learning_rate": 8.263264714409973e-08, + "loss": 0.1212, + "step": 13740 + }, + { + "epoch": 1.9933674024138306, + "grad_norm": 2.25, + "learning_rate": 6.8135691504784e-08, + "loss": 0.1476, + "step": 13750 + }, + { + "epoch": 1.9948171505200971, + "grad_norm": 1.3984375, + "learning_rate": 5.3638735865468255e-08, + "loss": 0.1515, + "step": 13760 + }, + { + "epoch": 1.9962668986263636, + "grad_norm": 1.9140625, + "learning_rate": 3.914178022615251e-08, + "loss": 0.1198, + "step": 13770 + }, + { + "epoch": 1.9977166467326302, + "grad_norm": 1.984375, + "learning_rate": 2.4644824586836768e-08, + "loss": 0.1503, + "step": 13780 + }, + { + "epoch": 1.9991663948388967, + "grad_norm": 1.390625, + "learning_rate": 1.014786894752102e-08, + "loss": 0.1505, + "step": 13790 + } + ], + "logging_steps": 10, + "max_steps": 13796, + "num_input_tokens_seen": 0, + "num_train_epochs": 2, + "save_steps": 500, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 5.952964652420825e+18, + "train_batch_size": 1, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-13796/training_args.bin b/checkpoint-13796/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..06838b874d6df8d2cf7fbec1647b56bbeee05efb --- /dev/null +++ b/checkpoint-13796/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0b692ac805c6882ecc47c1ef2d34907d8e095f209448815d8db422f0c5955b6d +size 5777 diff --git a/config.json b/config.json new file mode 100644 index 0000000000000000000000000000000000000000..bc35575d2212a40a1ecaf26ea55c3bdf0649ddfb --- /dev/null +++ b/config.json @@ -0,0 +1,31 @@ +{ + "architectures": [ + "GlmForCausalLM" + ], + "attention_bias": true, + "attention_dropout": 0.0, + "eos_token_id": [ + 151329, + 151336, + 151338 + ], + "head_dim": 128, + "hidden_act": "silu", + "hidden_size": 4096, + "initializer_range": 0.02, + "intermediate_size": 13696, + "max_position_embeddings": 131072, + "model_type": "glm", + "num_attention_heads": 32, + "num_hidden_layers": 40, + "num_key_value_heads": 2, + "pad_token_id": 151329, + "partial_rotary_factor": 0.5, + "rms_norm_eps": 1.5625e-07, + "rope_theta": 10000.0, + "tie_word_embeddings": false, + "torch_dtype": "bfloat16", + "transformers_version": "4.53.1", + "use_cache": true, + "vocab_size": 151552 +} diff --git a/generation_config.json b/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..d5a9c67e4e2fe314feb0a9b23144f0efeb644787 --- /dev/null +++ b/generation_config.json @@ -0,0 +1,10 @@ +{ + "_from_model_config": true, + "eos_token_id": [ + 151329, + 151336, + 151338 + ], + "pad_token_id": 151329, + "transformers_version": "4.53.1" +} diff --git a/model-00001-of-00004.safetensors b/model-00001-of-00004.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..93a5d3c0a5be2622b6df62300652b2f894f37c42 --- /dev/null +++ b/model-00001-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e7bd2f01ebc6d2da94caf7aea76f8eab8195e2b117d4603b9699873e797d7fb6 +size 4984133600 diff --git a/model-00002-of-00004.safetensors b/model-00002-of-00004.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..fc98b84b0f96cb86093ead554d72b5cd17669bc8 --- /dev/null +++ b/model-00002-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dc05900c7426c1fe59082ee18756c81947006dbb196af59fe4ff72aa3afeaa3a +size 4895075168 diff --git a/model-00003-of-00004.safetensors b/model-00003-of-00004.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d0678031174e9a6a5a57fb995119c8a184a8ec90 --- /dev/null +++ b/model-00003-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:06a9ec92a94d4ca828edf1470af8ff501f731f436ad1f5d839a2ea81f2de8f0f +size 4895075184 diff --git a/model-00004-of-00004.safetensors b/model-00004-of-00004.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..acef71bd5d941fabce50cb1bfc8410ebe3df7722 --- /dev/null +++ b/model-00004-of-00004.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a608f3679c3638d21306d4e413ce5441a97a65693fe9083193477b63a0eaba99 +size 4025669744 diff --git a/model.safetensors.index.json b/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..63641162b53de18c620d9bdad367eb87482d7fbf --- /dev/null +++ b/model.safetensors.index.json @@ -0,0 +1,451 @@ +{ + "metadata": { + "total_parameters": 9399951360, + "total_size": 18799902720 + }, + "weight_map": { + "lm_head.weight": "model-00004-of-00004.safetensors", + "model.embed_tokens.weight": "model-00001-of-00004.safetensors", + "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.20.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.20.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.21.self_attn.k_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.q_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.v_proj.bias": "model-00002-of-00004.safetensors", + "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.30.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.31.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.input_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.32.mlp.down_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.mlp.gate_up_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.post_attention_layernorm.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.32.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.33.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.33.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.33.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.33.self_attn.k_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.k_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.o_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.q_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.q_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.v_proj.bias": "model-00003-of-00004.safetensors", + "model.layers.33.self_attn.v_proj.weight": "model-00003-of-00004.safetensors", + "model.layers.34.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.34.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.34.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.34.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.34.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.35.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.35.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.36.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.36.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.37.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.37.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.38.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.38.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.input_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.39.mlp.down_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.mlp.gate_up_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.post_attention_layernorm.weight": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.k_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.k_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.o_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.q_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.q_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.v_proj.bias": "model-00004-of-00004.safetensors", + "model.layers.39.self_attn.v_proj.weight": "model-00004-of-00004.safetensors", + "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.input_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.down_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.mlp.gate_up_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.mlp.gate_up_proj.weight": "model-00002-of-00004.safetensors", + "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors", + "model.layers.9.self_attn.k_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.q_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.v_proj.bias": "model-00001-of-00004.safetensors", + "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00004.safetensors", + "model.norm.weight": "model-00004-of-00004.safetensors" + } +} diff --git a/runs/Jul08_19-12-12_192-222-50-4/events.out.tfevents.1752001932.192-222-50-4.11827.0 b/runs/Jul08_19-12-12_192-222-50-4/events.out.tfevents.1752001932.192-222-50-4.11827.0 new file mode 100644 index 0000000000000000000000000000000000000000..73ed23a4e880ed05271c824515ef92449595ed9a --- /dev/null +++ b/runs/Jul08_19-12-12_192-222-50-4/events.out.tfevents.1752001932.192-222-50-4.11827.0 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e4ba7c651c7522f690939d3b6b1d1961466c24d4824344661f5bbbf0927b0bd1 +size 6061 diff --git a/runs/Jul08_19-14-58_192-222-50-4/events.out.tfevents.1752002099.192-222-50-4.12776.0 b/runs/Jul08_19-14-58_192-222-50-4/events.out.tfevents.1752002099.192-222-50-4.12776.0 new file mode 100644 index 0000000000000000000000000000000000000000..8cbc8966042bcbe04192413f98e125216f15fb9b --- /dev/null +++ b/runs/Jul08_19-14-58_192-222-50-4/events.out.tfevents.1752002099.192-222-50-4.12776.0 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b1d6a3ab545865e5fc9bf4df45da3487f511e65e46ccb81a9fa22fba8ae38f8d +size 6475 diff --git a/runs/Jul08_19-20-58_192-222-50-4/events.out.tfevents.1752002459.192-222-50-4.13587.0 b/runs/Jul08_19-20-58_192-222-50-4/events.out.tfevents.1752002459.192-222-50-4.13587.0 new file mode 100644 index 0000000000000000000000000000000000000000..f619a66c0fc12a1079404c18315da935a7c54755 --- /dev/null +++ b/runs/Jul08_19-20-58_192-222-50-4/events.out.tfevents.1752002459.192-222-50-4.13587.0 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:26615887deea93211a95711d3f3bfe2edfa843212d3d868b4e225b132dd5f01b +size 6475 diff --git a/runs/Jul08_19-40-33_192-222-50-4/events.out.tfevents.1752003633.192-222-50-4.15226.0 b/runs/Jul08_19-40-33_192-222-50-4/events.out.tfevents.1752003633.192-222-50-4.15226.0 new file mode 100644 index 0000000000000000000000000000000000000000..433d05cb95a8a0871ba2ea47dad325913631c616 --- /dev/null +++ b/runs/Jul08_19-40-33_192-222-50-4/events.out.tfevents.1752003633.192-222-50-4.15226.0 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6621563a7bea52be27c7a7407052c0b5d58259abc89d4bacd17873d8fdd65ee9 +size 5026 diff --git a/runs/Jul08_19-41-29_192-222-50-4/events.out.tfevents.1752003689.192-222-50-4.15350.0 b/runs/Jul08_19-41-29_192-222-50-4/events.out.tfevents.1752003689.192-222-50-4.15350.0 new file mode 100644 index 0000000000000000000000000000000000000000..b6cb8b80c28976a5e966e0c1b9eea061518984ff --- /dev/null +++ b/runs/Jul08_19-41-29_192-222-50-4/events.out.tfevents.1752003689.192-222-50-4.15350.0 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5fe6b7802226a5ae5248f55c7bb5a0fa87bb7f298f6030df6984f3256d516f0b +size 52807 diff --git a/runs/Jul08_20-55-12_192-222-50-4/events.out.tfevents.1752008112.192-222-50-4.28917.0 b/runs/Jul08_20-55-12_192-222-50-4/events.out.tfevents.1752008112.192-222-50-4.28917.0 new file mode 100644 index 0000000000000000000000000000000000000000..8cb7b60d276a457e6501f325fc54ed508b15d78a --- /dev/null +++ b/runs/Jul08_20-55-12_192-222-50-4/events.out.tfevents.1752008112.192-222-50-4.28917.0 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a2edee08b569b354ede9ebe2dbedd14c07b20611206f824cc482280bf4b2be84 +size 5854 diff --git a/runs/Jul08_21-07-15_192-222-50-4/events.out.tfevents.1752008835.192-222-50-4.30219.0 b/runs/Jul08_21-07-15_192-222-50-4/events.out.tfevents.1752008835.192-222-50-4.30219.0 new file mode 100644 index 0000000000000000000000000000000000000000..773f8231ab0192e681d06517f9d1bde7a5f83463 --- /dev/null +++ b/runs/Jul08_21-07-15_192-222-50-4/events.out.tfevents.1752008835.192-222-50-4.30219.0 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3f2c271b064e33bbf90a33fef8a75b3598c162d07814d4638bef5283003de311 +size 47532 diff --git a/runs/Jul08_22-23-17_192-222-50-4/events.out.tfevents.1752013398.192-222-50-4.33656.0 b/runs/Jul08_22-23-17_192-222-50-4/events.out.tfevents.1752013398.192-222-50-4.33656.0 new file mode 100644 index 0000000000000000000000000000000000000000..83465f1d77ea99fb83454e708d94c1aaca62fa22 --- /dev/null +++ b/runs/Jul08_22-23-17_192-222-50-4/events.out.tfevents.1752013398.192-222-50-4.33656.0 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bebd75a4ccac79bd9be4ce4108b31ed391fccb833f2681872b9b034179fc8673 +size 4184 diff --git a/runs/Jul08_22-23-36_192-222-50-4/events.out.tfevents.1752013416.192-222-50-4.33741.0 b/runs/Jul08_22-23-36_192-222-50-4/events.out.tfevents.1752013416.192-222-50-4.33741.0 new file mode 100644 index 0000000000000000000000000000000000000000..292117f68688b8651e539ca271f8fd078eed70c7 --- /dev/null +++ b/runs/Jul08_22-23-36_192-222-50-4/events.out.tfevents.1752013416.192-222-50-4.33741.0 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6f90065756bc0c98ef5d7c2724dbe65c43376ae65fa03df7672c8fcc17b67f2e +size 296297 diff --git a/special_tokens_map.json b/special_tokens_map.json new file mode 100644 index 0000000000000000000000000000000000000000..582da4ad333feaaa0c427b4678a3e48447ea932f --- /dev/null +++ b/special_tokens_map.json @@ -0,0 +1,32 @@ +{ + "additional_special_tokens": [ + "<|endoftext|>", + "[MASK]", + "[gMASK]", + "[sMASK]", + "", + "", + "<|system|>", + "<|user|>", + "<|assistant|>", + "<|observation|>", + "<|begin_of_image|>", + "<|end_of_image|>", + "<|begin_of_video|>", + "<|end_of_video|>" + ], + "eos_token": { + "content": "<|endoftext|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + }, + "pad_token": { + "content": "<|endoftext|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false + } +} diff --git a/tokenizer.json b/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..2c24a7b2180f4b088af2671a705217ba4302fee5 --- /dev/null +++ b/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8a7269d6daa6328de533def0082ff9d3a825bb89036cbcc665c324f941f67fbf +size 19967863 diff --git a/tokenizer_config.json b/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..38214260739a871e4c13601f88d904a2ad1fd601 --- /dev/null +++ b/tokenizer_config.json @@ -0,0 +1,145 @@ +{ + "added_tokens_decoder": { + "151329": { + "content": "<|endoftext|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151330": { + "content": "[MASK]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151331": { + "content": "[gMASK]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151332": { + "content": "[sMASK]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151333": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151334": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151335": { + "content": "<|system|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151336": { + "content": "<|user|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151337": { + "content": "<|assistant|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151338": { + "content": "<|observation|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151339": { + "content": "<|begin_of_image|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151340": { + "content": "<|end_of_image|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151341": { + "content": "<|begin_of_video|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "151342": { + "content": "<|end_of_video|>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + } + }, + "additional_special_tokens": [ + "<|endoftext|>", + "[MASK]", + "[gMASK]", + "[sMASK]", + "", + "", + "<|system|>", + "<|user|>", + "<|assistant|>", + "<|observation|>", + "<|begin_of_image|>", + "<|end_of_image|>", + "<|begin_of_video|>", + "<|end_of_video|>" + ], + "clean_up_tokenization_spaces": false, + "do_lower_case": false, + "eos_token": "<|endoftext|>", + "extra_special_tokens": {}, + "model_input_names": [ + "input_ids", + "attention_mask" + ], + "model_max_length": 128000, + "pad_token": "<|endoftext|>", + "padding_side": "left", + "remove_space": false, + "tokenizer_class": "PreTrainedTokenizerFast" +} diff --git a/training_args.bin b/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..06838b874d6df8d2cf7fbec1647b56bbeee05efb --- /dev/null +++ b/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0b692ac805c6882ecc47c1ef2d34907d8e095f209448815d8db422f0c5955b6d +size 5777