Hugging Face's logo

Lansechen
/
deepseek-v2-lite-16b-chat-R1-Distill-batch8-numinamath

Text Generation
Transformers
Safetensors
deepseek_v2
Generated from Trainer
open-r1
trl
sft
conversational
custom_code
text-generation-inference
Model card Files Community
Lansechen commited on
Commit
754bca2
·
verified ·
1 Parent(s): 2e050a9

Model save

Browse files
Files changed (18) hide show
  1. README.md +58 -0
  2. all_results.json +8 -0
  3. config.json +61 -0
  4. generation_config.json +9 -0
  5. model-00001-of-00007.safetensors +3 -0
  6. model-00002-of-00007.safetensors +3 -0
  7. model-00003-of-00007.safetensors +3 -0
  8. model-00004-of-00007.safetensors +3 -0
  9. model-00005-of-00007.safetensors +3 -0
  10. model-00006-of-00007.safetensors +3 -0
  11. model-00007-of-00007.safetensors +3 -0
  12. model.safetensors.index.json +0 -0
  13. special_tokens_map.json +17 -0
  14. tokenizer.json +0 -0
  15. tokenizer_config.json +35 -0
  16. train_results.json +8 -0
  17. trainer_state.json +1411 -0
  18. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: deepseek-ai/DeepSeek-V2-Lite-Chat
3
+ library_name: transformers
4
+ model_name: deepseek-v2-lite-16b-chat-R1-Distill-batch8-numinamath
5
+ tags:
6
+ - generated_from_trainer
7
+ - trl
8
+ - sft
9
+ licence: license
10
+ ---
11
+
12
+ # Model Card for deepseek-v2-lite-16b-chat-R1-Distill-batch8-numinamath
13
+
14
+ This model is a fine-tuned version of [deepseek-ai/DeepSeek-V2-Lite-Chat](https://huggingface.co/deepseek-ai/DeepSeek-V2-Lite-Chat).
15
+ It has been trained using [TRL](https://github.com/huggingface/trl).
16
+
17
+ ## Quick start
18
+
19
+ ```python
20
+ from transformers import pipeline
21
+
22
+ question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
23
+ generator = pipeline("text-generation", model="Lansechen/deepseek-v2-lite-16b-chat-R1-Distill-batch8-numinamath", device="cuda")
24
+ output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
25
+ print(output["generated_text"])
26
+ ```
27
+
28
+ ## Training procedure
29
+
30
+ [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/chenran1995-the-chinese-university-of-hong-kong/huggingface/runs/zi06563j)
31
+
32
+
33
+ This model was trained with SFT.
34
+
35
+ ### Framework versions
36
+
37
+ - TRL: 0.15.0.dev0
38
+ - Transformers: 4.49.0.dev0
39
+ - Pytorch: 2.5.1+cu121
40
+ - Datasets: 3.2.0
41
+ - Tokenizers: 0.21.0
42
+
43
+ ## Citations
44
+
45
+
46
+
47
+ Cite TRL as:
48
+
49
+ ```bibtex
50
+ @misc{vonwerra2022trl,
51
+ title = {{TRL: Transformer Reinforcement Learning}},
52
+ author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
53
+ year = 2020,
54
+ journal = {GitHub repository},
55
+ publisher = {GitHub},
56
+ howpublished = {\url{https://github.com/huggingface/trl}}
57
+ }
58
+ ```
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "total_flos": 42703635087360.0,
3
+ "train_loss": 0.504338482067388,
4
+ "train_runtime": 61692.734,
5
+ "train_samples": 72441,
6
+ "train_samples_per_second": 0.446,
7
+ "train_steps_per_second": 0.014
8
+ }
config.json ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "deepseek-ai/DeepSeek-V2-Lite-Chat",
3
+ "architectures": [
4
+ "DeepseekV2ForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "auto_map": {
9
+ "AutoConfig": "deepseek-ai/DeepSeek-V2-Lite-Chat--configuration_deepseek.DeepseekV2Config",
10
+ "AutoModel": "deepseek-ai/DeepSeek-V2-Lite-Chat--modeling_deepseek.DeepseekV2Model",
11
+ "AutoModelForCausalLM": "deepseek-ai/DeepSeek-V2-Lite-Chat--modeling_deepseek.DeepseekV2ForCausalLM"
12
+ },
13
+ "aux_loss_alpha": 0.001,
14
+ "bos_token_id": 100000,
15
+ "eos_token_id": 100001,
16
+ "ep_size": 1,
17
+ "first_k_dense_replace": 1,
18
+ "hidden_act": "silu",
19
+ "hidden_size": 2048,
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 10944,
22
+ "kv_lora_rank": 512,
23
+ "max_position_embeddings": 163840,
24
+ "model_type": "deepseek_v2",
25
+ "moe_intermediate_size": 1408,
26
+ "moe_layer_freq": 1,
27
+ "n_group": 1,
28
+ "n_routed_experts": 64,
29
+ "n_shared_experts": 2,
30
+ "norm_topk_prob": false,
31
+ "num_attention_heads": 16,
32
+ "num_experts_per_tok": 6,
33
+ "num_hidden_layers": 27,
34
+ "num_key_value_heads": 16,
35
+ "pretraining_tp": 1,
36
+ "q_lora_rank": null,
37
+ "qk_nope_head_dim": 128,
38
+ "qk_rope_head_dim": 64,
39
+ "rms_norm_eps": 1e-06,
40
+ "rope_scaling": {
41
+ "beta_fast": 32,
42
+ "beta_slow": 1,
43
+ "factor": 40,
44
+ "mscale": 0.707,
45
+ "mscale_all_dim": 0.707,
46
+ "original_max_position_embeddings": 4096,
47
+ "type": "yarn"
48
+ },
49
+ "rope_theta": 10000,
50
+ "routed_scaling_factor": 1.0,
51
+ "scoring_func": "softmax",
52
+ "seq_aux": true,
53
+ "tie_word_embeddings": false,
54
+ "topk_group": 1,
55
+ "topk_method": "greedy",
56
+ "torch_dtype": "bfloat16",
57
+ "transformers_version": "4.49.0.dev0",
58
+ "use_cache": false,
59
+ "v_head_dim": 128,
60
+ "vocab_size": 102400
61
+ }
generation_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 100000,
4
+ "do_sample": true,
5
+ "eos_token_id": 100001,
6
+ "temperature": 0.3,
7
+ "top_p": 0.95,
8
+ "transformers_version": "4.49.0.dev0"
9
+ }
model-00001-of-00007.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:57e594d30c2dfa387fcdbe2d237e8ee4a3c1488125c065268b7a37bbe34e9184
3
+ size 4994763632
model-00002-of-00007.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4d32d77810429c5167f35251a87213c84461e71ba3b18d5704b9038d03b5062c
3
+ size 4995044944
model-00003-of-00007.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:96188d7d09d39d6abad6038cc1fd4f7df7cd60088dc388c4e1e2f39219d90464
3
+ size 4996085000
model-00004-of-00007.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:124ce27c130daf1efc251f2898fcb38ae56791ca92160e9e99c1889bd90d56b3
3
+ size 4996085224
model-00005-of-00007.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cfbc0c405629ef2e9e5e5bfb8ef06351e458bccf452ebeefabb2af050da6ea3e
3
+ size 4996085224
model-00006-of-00007.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:864a28cfcb59e99a8dd19fea7c01a09386bc48b08939da28fc6706403b632467
3
+ size 4995045792
model-00007-of-00007.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:391f67c0f862e41d4bcf87796507ec8064926e590f0fe1e550e21d1e1171e01b
3
+ size 1440515736
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
special_tokens_map.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<｜begin▁of▁sentence｜>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<｜end▁of▁sentence｜>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "<｜end▁of▁sentence｜>"
17
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": null,
5
+ "added_tokens_decoder": {
6
+ "100000": {
7
+ "content": "<｜begin▁of▁sentence｜>",
8
+ "lstrip": false,
9
+ "normalized": true,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "100001": {
15
+ "content": "<｜end▁of▁sentence｜>",
16
+ "lstrip": false,
17
+ "normalized": true,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": true
21
+ }
22
+ },
23
+ "bos_token": "<｜begin▁of▁sentence｜>",
24
+ "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{{ bos_token }}{% for message in messages %}{% if message['role'] == 'user' %}{{ 'User: ' + message['content'] + '\n\n' }}{% elif message['role'] == 'assistant' %}{{ 'Assistant: ' + message['content'] + eos_token }}{% elif message['role'] == 'system' %}{{ message['content'] + '\n\n' }}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}",
25
+ "clean_up_tokenization_spaces": false,
26
+ "eos_token": "<｜end▁of▁sentence｜>",
27
+ "extra_special_tokens": {},
28
+ "legacy": true,
29
+ "model_max_length": 16384,
30
+ "pad_token": "<｜end▁of▁sentence｜>",
31
+ "sp_model_kwargs": {},
32
+ "tokenizer_class": "LlamaTokenizer",
33
+ "unk_token": null,
34
+ "use_default_system_prompt": false
35
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "total_flos": 42703635087360.0,
3
+ "train_loss": 0.504338482067388,
4
+ "train_runtime": 61692.734,
5
+ "train_samples": 72441,
6
+ "train_samples_per_second": 0.446,
7
+ "train_steps_per_second": 0.014
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,1411 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.9997090485888857,
5
+ "eval_steps": 100,
6
+ "global_step": 859,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.005819028222286878,
13
+ "grad_norm": 4.729677200317383,
14
+ "learning_rate": 1.1627906976744188e-06,
15
+ "loss": 0.8687,
16
+ "mean_token_accuracy": 0.8001353883684503,
17
+ "step": 5
18
+ },
19
+ {
20
+ "epoch": 0.011638056444573755,
21
+ "grad_norm": 3.4830121994018555,
22
+ "learning_rate": 2.3255813953488376e-06,
23
+ "loss": 0.8575,
24
+ "mean_token_accuracy": 0.7992099680535122,
25
+ "step": 10
26
+ },
27
+ {
28
+ "epoch": 0.017457084666860634,
29
+ "grad_norm": 1.5148690938949585,
30
+ "learning_rate": 3.4883720930232564e-06,
31
+ "loss": 0.7863,
32
+ "mean_token_accuracy": 0.8034960847656469,
33
+ "step": 15
34
+ },
35
+ {
36
+ "epoch": 0.02327611288914751,
37
+ "grad_norm": 1.5541874170303345,
38
+ "learning_rate": 4.651162790697675e-06,
39
+ "loss": 0.7052,
40
+ "mean_token_accuracy": 0.813427774762903,
41
+ "step": 20
42
+ },
43
+ {
44
+ "epoch": 0.02909514111143439,
45
+ "grad_norm": 1.0066415071487427,
46
+ "learning_rate": 5.8139534883720935e-06,
47
+ "loss": 0.6611,
48
+ "mean_token_accuracy": 0.8186242834655136,
49
+ "step": 25
50
+ },
51
+ {
52
+ "epoch": 0.03491416933372127,
53
+ "grad_norm": 0.7388224601745605,
54
+ "learning_rate": 6.976744186046513e-06,
55
+ "loss": 0.6075,
56
+ "mean_token_accuracy": 0.8315073685463057,
57
+ "step": 30
58
+ },
59
+ {
60
+ "epoch": 0.04073319755600815,
61
+ "grad_norm": 0.5410726070404053,
62
+ "learning_rate": 8.139534883720931e-06,
63
+ "loss": 0.5962,
64
+ "mean_token_accuracy": 0.832317557757456,
65
+ "step": 35
66
+ },
67
+ {
68
+ "epoch": 0.04655222577829502,
69
+ "grad_norm": 0.53746497631073,
70
+ "learning_rate": 9.30232558139535e-06,
71
+ "loss": 0.5733,
72
+ "mean_token_accuracy": 0.8373400698056097,
73
+ "step": 40
74
+ },
75
+ {
76
+ "epoch": 0.0523712540005819,
77
+ "grad_norm": 0.512347400188446,
78
+ "learning_rate": 1.046511627906977e-05,
79
+ "loss": 0.5627,
80
+ "mean_token_accuracy": 0.8384384618000557,
81
+ "step": 45
82
+ },
83
+ {
84
+ "epoch": 0.05819028222286878,
85
+ "grad_norm": 0.5112928748130798,
86
+ "learning_rate": 1.1627906976744187e-05,
87
+ "loss": 0.5752,
88
+ "mean_token_accuracy": 0.836557222748716,
89
+ "step": 50
90
+ },
91
+ {
92
+ "epoch": 0.06400931044515566,
93
+ "grad_norm": 0.4823078513145447,
94
+ "learning_rate": 1.2790697674418606e-05,
95
+ "loss": 0.5546,
96
+ "mean_token_accuracy": 0.8406492042617113,
97
+ "step": 55
98
+ },
99
+ {
100
+ "epoch": 0.06982833866744254,
101
+ "grad_norm": 0.48803916573524475,
102
+ "learning_rate": 1.3953488372093025e-05,
103
+ "loss": 0.5552,
104
+ "mean_token_accuracy": 0.8404829421801073,
105
+ "step": 60
106
+ },
107
+ {
108
+ "epoch": 0.07564736688972941,
109
+ "grad_norm": 0.5077307224273682,
110
+ "learning_rate": 1.5116279069767443e-05,
111
+ "loss": 0.5537,
112
+ "mean_token_accuracy": 0.8408575111390093,
113
+ "step": 65
114
+ },
115
+ {
116
+ "epoch": 0.0814663951120163,
117
+ "grad_norm": 0.5025640726089478,
118
+ "learning_rate": 1.6279069767441862e-05,
119
+ "loss": 0.5324,
120
+ "mean_token_accuracy": 0.8454815162857129,
121
+ "step": 70
122
+ },
123
+ {
124
+ "epoch": 0.08728542333430317,
125
+ "grad_norm": 0.4802047610282898,
126
+ "learning_rate": 1.744186046511628e-05,
127
+ "loss": 0.5298,
128
+ "mean_token_accuracy": 0.8467988429180437,
129
+ "step": 75
130
+ },
131
+ {
132
+ "epoch": 0.09310445155659004,
133
+ "grad_norm": 0.495412141084671,
134
+ "learning_rate": 1.86046511627907e-05,
135
+ "loss": 0.5332,
136
+ "mean_token_accuracy": 0.847131608146347,
137
+ "step": 80
138
+ },
139
+ {
140
+ "epoch": 0.09892347977887693,
141
+ "grad_norm": 0.4867161512374878,
142
+ "learning_rate": 1.9767441860465116e-05,
143
+ "loss": 0.5315,
144
+ "mean_token_accuracy": 0.8456074647350855,
145
+ "step": 85
146
+ },
147
+ {
148
+ "epoch": 0.1047425080011638,
149
+ "grad_norm": 0.5120564699172974,
150
+ "learning_rate": 1.999867863992634e-05,
151
+ "loss": 0.5419,
152
+ "mean_token_accuracy": 0.8431668032243257,
153
+ "step": 90
154
+ },
155
+ {
156
+ "epoch": 0.11056153622345069,
157
+ "grad_norm": 0.5223878026008606,
158
+ "learning_rate": 1.9993311213101313e-05,
159
+ "loss": 0.5379,
160
+ "mean_token_accuracy": 0.8446533035042771,
161
+ "step": 95
162
+ },
163
+ {
164
+ "epoch": 0.11638056444573756,
165
+ "grad_norm": 0.4827958643436432,
166
+ "learning_rate": 1.998381734913258e-05,
167
+ "loss": 0.528,
168
+ "mean_token_accuracy": 0.8463738754778392,
169
+ "step": 100
170
+ },
171
+ {
172
+ "epoch": 0.12219959266802444,
173
+ "grad_norm": 0.49347934126853943,
174
+ "learning_rate": 1.997020096822557e-05,
175
+ "loss": 0.524,
176
+ "mean_token_accuracy": 0.8469525801009745,
177
+ "step": 105
178
+ },
179
+ {
180
+ "epoch": 0.12801862089031132,
181
+ "grad_norm": 0.4837639033794403,
182
+ "learning_rate": 1.9952467692855043e-05,
183
+ "loss": 0.5411,
184
+ "mean_token_accuracy": 0.8443662513099179,
185
+ "step": 110
186
+ },
187
+ {
188
+ "epoch": 0.13383764911259818,
189
+ "grad_norm": 0.44964534044265747,
190
+ "learning_rate": 1.993062484544341e-05,
191
+ "loss": 0.5122,
192
+ "mean_token_accuracy": 0.8500057242518922,
193
+ "step": 115
194
+ },
195
+ {
196
+ "epoch": 0.13965667733488507,
197
+ "grad_norm": 0.4763283133506775,
198
+ "learning_rate": 1.990468144533722e-05,
199
+ "loss": 0.5467,
200
+ "mean_token_accuracy": 0.842521201317496,
201
+ "step": 120
202
+ },
203
+ {
204
+ "epoch": 0.14547570555717196,
205
+ "grad_norm": 0.4805982708930969,
206
+ "learning_rate": 1.9874648205082847e-05,
207
+ "loss": 0.5267,
208
+ "mean_token_accuracy": 0.8467973214319761,
209
+ "step": 125
210
+ },
211
+ {
212
+ "epoch": 0.15129473377945882,
213
+ "grad_norm": 0.47071903944015503,
214
+ "learning_rate": 1.9840537526003085e-05,
215
+ "loss": 0.534,
216
+ "mean_token_accuracy": 0.8436821750241809,
217
+ "step": 130
218
+ },
219
+ {
220
+ "epoch": 0.1571137620017457,
221
+ "grad_norm": 0.49743327498435974,
222
+ "learning_rate": 1.9802363493076392e-05,
223
+ "loss": 0.5175,
224
+ "mean_token_accuracy": 0.8488135271716413,
225
+ "step": 135
226
+ },
227
+ {
228
+ "epoch": 0.1629327902240326,
229
+ "grad_norm": 0.46099644899368286,
230
+ "learning_rate": 1.9760141869120917e-05,
231
+ "loss": 0.5188,
232
+ "mean_token_accuracy": 0.8494274636540204,
233
+ "step": 140
234
+ },
235
+ {
236
+ "epoch": 0.16875181844631945,
237
+ "grad_norm": 0.43989109992980957,
238
+ "learning_rate": 1.971389008828573e-05,
239
+ "loss": 0.51,
240
+ "mean_token_accuracy": 0.8514118628602676,
241
+ "step": 145
242
+ },
243
+ {
244
+ "epoch": 0.17457084666860634,
245
+ "grad_norm": 0.44857239723205566,
246
+ "learning_rate": 1.9663627248851903e-05,
247
+ "loss": 0.505,
248
+ "mean_token_accuracy": 0.8521678886513273,
249
+ "step": 150
250
+ },
251
+ {
252
+ "epoch": 0.18038987489089323,
253
+ "grad_norm": 0.4434855580329895,
254
+ "learning_rate": 1.9609374105346458e-05,
255
+ "loss": 0.5348,
256
+ "mean_token_accuracy": 0.8445110785073409,
257
+ "step": 155
258
+ },
259
+ {
260
+ "epoch": 0.18620890311318009,
261
+ "grad_norm": 0.45866096019744873,
262
+ "learning_rate": 1.9551153059972397e-05,
263
+ "loss": 0.5387,
264
+ "mean_token_accuracy": 0.8436002910675257,
265
+ "step": 160
266
+ },
267
+ {
268
+ "epoch": 0.19202793133546697,
269
+ "grad_norm": 0.4627978801727295,
270
+ "learning_rate": 1.948898815335841e-05,
271
+ "loss": 0.5164,
272
+ "mean_token_accuracy": 0.848416589531312,
273
+ "step": 165
274
+ },
275
+ {
276
+ "epoch": 0.19784695955775386,
277
+ "grad_norm": 0.4659474790096283,
278
+ "learning_rate": 1.9422905054631996e-05,
279
+ "loss": 0.5129,
280
+ "mean_token_accuracy": 0.8507672546269303,
281
+ "step": 170
282
+ },
283
+ {
284
+ "epoch": 0.20366598778004075,
285
+ "grad_norm": 0.4798222482204437,
286
+ "learning_rate": 1.9352931050820175e-05,
287
+ "loss": 0.5173,
288
+ "mean_token_accuracy": 0.8474816378041764,
289
+ "step": 175
290
+ },
291
+ {
292
+ "epoch": 0.2094850160023276,
293
+ "grad_norm": 0.5056359171867371,
294
+ "learning_rate": 1.9279095035582153e-05,
295
+ "loss": 0.5128,
296
+ "mean_token_accuracy": 0.8497545293493058,
297
+ "step": 180
298
+ },
299
+ {
300
+ "epoch": 0.2153040442246145,
301
+ "grad_norm": 0.41810187697410583,
302
+ "learning_rate": 1.9201427497278518e-05,
303
+ "loss": 0.5105,
304
+ "mean_token_accuracy": 0.8495568220795893,
305
+ "step": 185
306
+ },
307
+ {
308
+ "epoch": 0.22112307244690138,
309
+ "grad_norm": 0.45786747336387634,
310
+ "learning_rate": 1.911996050638202e-05,
311
+ "loss": 0.508,
312
+ "mean_token_accuracy": 0.8502620839044074,
313
+ "step": 190
314
+ },
315
+ {
316
+ "epoch": 0.22694210066918824,
317
+ "grad_norm": 0.4433153569698334,
318
+ "learning_rate": 1.9034727702235023e-05,
319
+ "loss": 0.5251,
320
+ "mean_token_accuracy": 0.8470169206232894,
321
+ "step": 195
322
+ },
323
+ {
324
+ "epoch": 0.23276112889147513,
325
+ "grad_norm": 0.43002447485923767,
326
+ "learning_rate": 1.8945764279159144e-05,
327
+ "loss": 0.51,
328
+ "mean_token_accuracy": 0.8515899255383441,
329
+ "step": 200
330
+ },
331
+ {
332
+ "epoch": 0.23858015711376201,
333
+ "grad_norm": 0.43477922677993774,
334
+ "learning_rate": 1.8853106971922833e-05,
335
+ "loss": 0.5162,
336
+ "mean_token_accuracy": 0.8489151326449708,
337
+ "step": 205
338
+ },
339
+ {
340
+ "epoch": 0.24439918533604887,
341
+ "grad_norm": 0.4386807978153229,
342
+ "learning_rate": 1.8756794040572834e-05,
343
+ "loss": 0.528,
344
+ "mean_token_accuracy": 0.8479445785479058,
345
+ "step": 210
346
+ },
347
+ {
348
+ "epoch": 0.25021821355833573,
349
+ "grad_norm": 0.43842801451683044,
350
+ "learning_rate": 1.8656865254635877e-05,
351
+ "loss": 0.5007,
352
+ "mean_token_accuracy": 0.852635679878258,
353
+ "step": 215
354
+ },
355
+ {
356
+ "epoch": 0.25603724178062265,
357
+ "grad_norm": 0.46626922488212585,
358
+ "learning_rate": 1.8553361876697025e-05,
359
+ "loss": 0.5274,
360
+ "mean_token_accuracy": 0.8479028685155747,
361
+ "step": 220
362
+ },
363
+ {
364
+ "epoch": 0.2618562700029095,
365
+ "grad_norm": 0.4394637644290924,
366
+ "learning_rate": 1.8446326645361542e-05,
367
+ "loss": 0.5113,
368
+ "mean_token_accuracy": 0.8496161991247855,
369
+ "step": 225
370
+ },
371
+ {
372
+ "epoch": 0.26767529822519637,
373
+ "grad_norm": 0.46903783082962036,
374
+ "learning_rate": 1.8335803757607274e-05,
375
+ "loss": 0.5042,
376
+ "mean_token_accuracy": 0.851652269825092,
377
+ "step": 230
378
+ },
379
+ {
380
+ "epoch": 0.2734943264474833,
381
+ "grad_norm": 0.4322178363800049,
382
+ "learning_rate": 1.822183885053483e-05,
383
+ "loss": 0.5254,
384
+ "mean_token_accuracy": 0.8445282817492037,
385
+ "step": 235
386
+ },
387
+ {
388
+ "epoch": 0.27931335466977014,
389
+ "grad_norm": 0.4331161677837372,
390
+ "learning_rate": 1.8104478982523117e-05,
391
+ "loss": 0.5043,
392
+ "mean_token_accuracy": 0.8515775238870337,
393
+ "step": 240
394
+ },
395
+ {
396
+ "epoch": 0.285132382892057,
397
+ "grad_norm": 0.4396701753139496,
398
+ "learning_rate": 1.7983772613798006e-05,
399
+ "loss": 0.5003,
400
+ "mean_token_accuracy": 0.8519097787097343,
401
+ "step": 245
402
+ },
403
+ {
404
+ "epoch": 0.2909514111143439,
405
+ "grad_norm": 0.4399448335170746,
406
+ "learning_rate": 1.7859769586422122e-05,
407
+ "loss": 0.495,
408
+ "mean_token_accuracy": 0.8534961898688037,
409
+ "step": 250
410
+ },
411
+ {
412
+ "epoch": 0.2967704393366308,
413
+ "grad_norm": 0.4380975067615509,
414
+ "learning_rate": 1.773252110371409e-05,
415
+ "loss": 0.5117,
416
+ "mean_token_accuracy": 0.849924917367521,
417
+ "step": 255
418
+ },
419
+ {
420
+ "epoch": 0.30258946755891764,
421
+ "grad_norm": 0.4397619962692261,
422
+ "learning_rate": 1.7602079709105644e-05,
423
+ "loss": 0.4925,
424
+ "mean_token_accuracy": 0.8539092350587474,
425
+ "step": 260
426
+ },
427
+ {
428
+ "epoch": 0.30840849578120455,
429
+ "grad_norm": 0.4179295003414154,
430
+ "learning_rate": 1.7468499264445405e-05,
431
+ "loss": 0.4958,
432
+ "mean_token_accuracy": 0.8536148051263105,
433
+ "step": 265
434
+ },
435
+ {
436
+ "epoch": 0.3142275240034914,
437
+ "grad_norm": 0.4244683086872101,
438
+ "learning_rate": 1.733183492775825e-05,
439
+ "loss": 0.5052,
440
+ "mean_token_accuracy": 0.8512700493372496,
441
+ "step": 270
442
+ },
443
+ {
444
+ "epoch": 0.32004655222577827,
445
+ "grad_norm": 0.44008669257164,
446
+ "learning_rate": 1.7192143130469466e-05,
447
+ "loss": 0.5043,
448
+ "mean_token_accuracy": 0.8509524474133208,
449
+ "step": 275
450
+ },
451
+ {
452
+ "epoch": 0.3258655804480652,
453
+ "grad_norm": 0.42162829637527466,
454
+ "learning_rate": 1.7049481554103107e-05,
455
+ "loss": 0.5036,
456
+ "mean_token_accuracy": 0.8519439016703739,
457
+ "step": 280
458
+ },
459
+ {
460
+ "epoch": 0.33168460867035204,
461
+ "grad_norm": 0.44394752383232117,
462
+ "learning_rate": 1.690390910646411e-05,
463
+ "loss": 0.4962,
464
+ "mean_token_accuracy": 0.8529925945929515,
465
+ "step": 285
466
+ },
467
+ {
468
+ "epoch": 0.3375036368926389,
469
+ "grad_norm": 0.43867236375808716,
470
+ "learning_rate": 1.6755485897314122e-05,
471
+ "loss": 0.4986,
472
+ "mean_token_accuracy": 0.8537370793513223,
473
+ "step": 290
474
+ },
475
+ {
476
+ "epoch": 0.3433226651149258,
477
+ "grad_norm": 0.4371063709259033,
478
+ "learning_rate": 1.6604273213550957e-05,
479
+ "loss": 0.4994,
480
+ "mean_token_accuracy": 0.853144893346005,
481
+ "step": 295
482
+ },
483
+ {
484
+ "epoch": 0.3491416933372127,
485
+ "grad_norm": 0.4125002920627594,
486
+ "learning_rate": 1.6450333493902046e-05,
487
+ "loss": 0.5145,
488
+ "mean_token_accuracy": 0.8495829518753062,
489
+ "step": 300
490
+ },
491
+ {
492
+ "epoch": 0.35496072155949954,
493
+ "grad_norm": 0.422319233417511,
494
+ "learning_rate": 1.6293730303142218e-05,
495
+ "loss": 0.5098,
496
+ "mean_token_accuracy": 0.8512014355945091,
497
+ "step": 305
498
+ },
499
+ {
500
+ "epoch": 0.36077974978178645,
501
+ "grad_norm": 0.4452224671840668,
502
+ "learning_rate": 1.6134528305846537e-05,
503
+ "loss": 0.4868,
504
+ "mean_token_accuracy": 0.856624218400263,
505
+ "step": 310
506
+ },
507
+ {
508
+ "epoch": 0.3665987780040733,
509
+ "grad_norm": 0.4240838885307312,
510
+ "learning_rate": 1.5972793239689038e-05,
511
+ "loss": 0.4885,
512
+ "mean_token_accuracy": 0.8560715988512444,
513
+ "step": 315
514
+ },
515
+ {
516
+ "epoch": 0.37241780622636017,
517
+ "grad_norm": 0.4351670742034912,
518
+ "learning_rate": 1.5808591888298314e-05,
519
+ "loss": 0.4949,
520
+ "mean_token_accuracy": 0.8541411151521825,
521
+ "step": 320
522
+ },
523
+ {
524
+ "epoch": 0.3782368344486471,
525
+ "grad_norm": 0.41906318068504333,
526
+ "learning_rate": 1.5641992053681213e-05,
527
+ "loss": 0.5087,
528
+ "mean_token_accuracy": 0.850579562045576,
529
+ "step": 325
530
+ },
531
+ {
532
+ "epoch": 0.38405586267093395,
533
+ "grad_norm": 0.42045196890830994,
534
+ "learning_rate": 1.5473062528226082e-05,
535
+ "loss": 0.4853,
536
+ "mean_token_accuracy": 0.8547947922524161,
537
+ "step": 330
538
+ },
539
+ {
540
+ "epoch": 0.3898748908932208,
541
+ "grad_norm": 0.43161919713020325,
542
+ "learning_rate": 1.530187306629697e-05,
543
+ "loss": 0.5078,
544
+ "mean_token_accuracy": 0.851508238713485,
545
+ "step": 335
546
+ },
547
+ {
548
+ "epoch": 0.3956939191155077,
549
+ "grad_norm": 0.4057161509990692,
550
+ "learning_rate": 1.5128494355430698e-05,
551
+ "loss": 0.4848,
552
+ "mean_token_accuracy": 0.8560923491214802,
553
+ "step": 340
554
+ },
555
+ {
556
+ "epoch": 0.4015129473377946,
557
+ "grad_norm": 0.43893197178840637,
558
+ "learning_rate": 1.4952997987148554e-05,
559
+ "loss": 0.5114,
560
+ "mean_token_accuracy": 0.849725545842043,
561
+ "step": 345
562
+ },
563
+ {
564
+ "epoch": 0.4073319755600815,
565
+ "grad_norm": 0.432678759098053,
566
+ "learning_rate": 1.4775456427394732e-05,
567
+ "loss": 0.5008,
568
+ "mean_token_accuracy": 0.8532302583194186,
569
+ "step": 350
570
+ },
571
+ {
572
+ "epoch": 0.41315100378236835,
573
+ "grad_norm": 0.40871691703796387,
574
+ "learning_rate": 1.4595942986613696e-05,
575
+ "loss": 0.4984,
576
+ "mean_token_accuracy": 0.8532044769095325,
577
+ "step": 355
578
+ },
579
+ {
580
+ "epoch": 0.4189700320046552,
581
+ "grad_norm": 0.43299028277397156,
582
+ "learning_rate": 1.4414531789478841e-05,
583
+ "loss": 0.5092,
584
+ "mean_token_accuracy": 0.8504414872190281,
585
+ "step": 360
586
+ },
587
+ {
588
+ "epoch": 0.42478906022694213,
589
+ "grad_norm": 0.41469907760620117,
590
+ "learning_rate": 1.423129774428495e-05,
591
+ "loss": 0.496,
592
+ "mean_token_accuracy": 0.8539978662641772,
593
+ "step": 365
594
+ },
595
+ {
596
+ "epoch": 0.430608088449229,
597
+ "grad_norm": 0.44283366203308105,
598
+ "learning_rate": 1.4046316512017044e-05,
599
+ "loss": 0.4887,
600
+ "mean_token_accuracy": 0.8548189689797085,
601
+ "step": 370
602
+ },
603
+ {
604
+ "epoch": 0.43642711667151585,
605
+ "grad_norm": 0.43634092807769775,
606
+ "learning_rate": 1.3859664475108483e-05,
607
+ "loss": 0.4949,
608
+ "mean_token_accuracy": 0.8529859121625627,
609
+ "step": 375
610
+ },
611
+ {
612
+ "epoch": 0.44224614489380276,
613
+ "grad_norm": 0.41822701692581177,
614
+ "learning_rate": 1.3671418705901116e-05,
615
+ "loss": 0.5025,
616
+ "mean_token_accuracy": 0.8515156968159479,
617
+ "step": 380
618
+ },
619
+ {
620
+ "epoch": 0.4480651731160896,
621
+ "grad_norm": 0.42008936405181885,
622
+ "learning_rate": 1.3481656934820588e-05,
623
+ "loss": 0.4957,
624
+ "mean_token_accuracy": 0.8534913304740183,
625
+ "step": 385
626
+ },
627
+ {
628
+ "epoch": 0.4538842013383765,
629
+ "grad_norm": 0.4045666754245758,
630
+ "learning_rate": 1.32904575182799e-05,
631
+ "loss": 0.4837,
632
+ "mean_token_accuracy": 0.8559156045618721,
633
+ "step": 390
634
+ },
635
+ {
636
+ "epoch": 0.4597032295606634,
637
+ "grad_norm": 0.4344455301761627,
638
+ "learning_rate": 1.309789940632448e-05,
639
+ "loss": 0.4863,
640
+ "mean_token_accuracy": 0.8565342245889443,
641
+ "step": 395
642
+ },
643
+ {
644
+ "epoch": 0.46552225778295026,
645
+ "grad_norm": 0.4040561616420746,
646
+ "learning_rate": 1.2904062110032145e-05,
647
+ "loss": 0.4888,
648
+ "mean_token_accuracy": 0.8559227694031748,
649
+ "step": 400
650
+ },
651
+ {
652
+ "epoch": 0.4713412860052371,
653
+ "grad_norm": 0.41261550784111023,
654
+ "learning_rate": 1.270902566868139e-05,
655
+ "loss": 0.4999,
656
+ "mean_token_accuracy": 0.8528450147039587,
657
+ "step": 405
658
+ },
659
+ {
660
+ "epoch": 0.47716031422752403,
661
+ "grad_norm": 0.41591835021972656,
662
+ "learning_rate": 1.2512870616701571e-05,
663
+ "loss": 0.4858,
664
+ "mean_token_accuracy": 0.8563491876605844,
665
+ "step": 410
666
+ },
667
+ {
668
+ "epoch": 0.4829793424498109,
669
+ "grad_norm": 0.4135805368423462,
670
+ "learning_rate": 1.2315677950418645e-05,
671
+ "loss": 0.491,
672
+ "mean_token_accuracy": 0.8547786818046786,
673
+ "step": 415
674
+ },
675
+ {
676
+ "epoch": 0.48879837067209775,
677
+ "grad_norm": 0.42353108525276184,
678
+ "learning_rate": 1.2117529094610177e-05,
679
+ "loss": 0.4889,
680
+ "mean_token_accuracy": 0.8551664301669952,
681
+ "step": 420
682
+ },
683
+ {
684
+ "epoch": 0.49461739889438466,
685
+ "grad_norm": 0.4162733554840088,
686
+ "learning_rate": 1.1918505868883414e-05,
687
+ "loss": 0.4838,
688
+ "mean_token_accuracy": 0.856896010365834,
689
+ "step": 425
690
+ },
691
+ {
692
+ "epoch": 0.5004364271166715,
693
+ "grad_norm": 0.41785794496536255,
694
+ "learning_rate": 1.1718690453890356e-05,
695
+ "loss": 0.4917,
696
+ "mean_token_accuracy": 0.8560505977625705,
697
+ "step": 430
698
+ },
699
+ {
700
+ "epoch": 0.5062554553389584,
701
+ "grad_norm": 0.44993114471435547,
702
+ "learning_rate": 1.1518165357393716e-05,
703
+ "loss": 0.5001,
704
+ "mean_token_accuracy": 0.852639218024868,
705
+ "step": 435
706
+ },
707
+ {
708
+ "epoch": 0.5120744835612453,
709
+ "grad_norm": 0.41796672344207764,
710
+ "learning_rate": 1.1317013380197832e-05,
711
+ "loss": 0.4823,
712
+ "mean_token_accuracy": 0.8560310313992504,
713
+ "step": 440
714
+ },
715
+ {
716
+ "epoch": 0.5178935117835322,
717
+ "grad_norm": 0.44025880098342896,
718
+ "learning_rate": 1.111531758195856e-05,
719
+ "loss": 0.48,
720
+ "mean_token_accuracy": 0.8575471706652287,
721
+ "step": 445
722
+ },
723
+ {
724
+ "epoch": 0.523712540005819,
725
+ "grad_norm": 0.4278118312358856,
726
+ "learning_rate": 1.091316124688629e-05,
727
+ "loss": 0.4971,
728
+ "mean_token_accuracy": 0.852513281695311,
729
+ "step": 450
730
+ },
731
+ {
732
+ "epoch": 0.5295315682281059,
733
+ "grad_norm": 0.42229601740837097,
734
+ "learning_rate": 1.0710627849356246e-05,
735
+ "loss": 0.4773,
736
+ "mean_token_accuracy": 0.8587941342459156,
737
+ "step": 455
738
+ },
739
+ {
740
+ "epoch": 0.5353505964503927,
741
+ "grad_norm": 0.4171890318393707,
742
+ "learning_rate": 1.0507801019440235e-05,
743
+ "loss": 0.4776,
744
+ "mean_token_accuracy": 0.8578564343102109,
745
+ "step": 460
746
+ },
747
+ {
748
+ "epoch": 0.5411696246726797,
749
+ "grad_norm": 0.41802269220352173,
750
+ "learning_rate": 1.0304764508374152e-05,
751
+ "loss": 0.4837,
752
+ "mean_token_accuracy": 0.8558379914367457,
753
+ "step": 465
754
+ },
755
+ {
756
+ "epoch": 0.5469886528949666,
757
+ "grad_norm": 0.43975040316581726,
758
+ "learning_rate": 1.0101602153975398e-05,
759
+ "loss": 0.4925,
760
+ "mean_token_accuracy": 0.8547882250775738,
761
+ "step": 470
762
+ },
763
+ {
764
+ "epoch": 0.5528076811172534,
765
+ "grad_norm": 0.3969772160053253,
766
+ "learning_rate": 9.898397846024604e-06,
767
+ "loss": 0.4725,
768
+ "mean_token_accuracy": 0.8595373324490415,
769
+ "step": 475
770
+ },
771
+ {
772
+ "epoch": 0.5586267093395403,
773
+ "grad_norm": 0.41007471084594727,
774
+ "learning_rate": 9.69523549162585e-06,
775
+ "loss": 0.4851,
776
+ "mean_token_accuracy": 0.855284327692498,
777
+ "step": 480
778
+ },
779
+ {
780
+ "epoch": 0.5644457375618271,
781
+ "grad_norm": 0.4255843758583069,
782
+ "learning_rate": 9.492198980559766e-06,
783
+ "loss": 0.4905,
784
+ "mean_token_accuracy": 0.8558610098037495,
785
+ "step": 485
786
+ },
787
+ {
788
+ "epoch": 0.570264765784114,
789
+ "grad_norm": 0.41996172070503235,
790
+ "learning_rate": 9.289372150643759e-06,
791
+ "loss": 0.4882,
792
+ "mean_token_accuracy": 0.8572012931662899,
793
+ "step": 490
794
+ },
795
+ {
796
+ "epoch": 0.576083794006401,
797
+ "grad_norm": 0.4162617623806,
798
+ "learning_rate": 9.08683875311371e-06,
799
+ "loss": 0.4932,
800
+ "mean_token_accuracy": 0.8533011128757575,
801
+ "step": 495
802
+ },
803
+ {
804
+ "epoch": 0.5819028222286878,
805
+ "grad_norm": 0.4218461811542511,
806
+ "learning_rate": 8.884682418041443e-06,
807
+ "loss": 0.4795,
808
+ "mean_token_accuracy": 0.8570308869685629,
809
+ "step": 500
810
+ },
811
+ {
812
+ "epoch": 0.5877218504509747,
813
+ "grad_norm": 0.41105833649635315,
814
+ "learning_rate": 8.682986619802171e-06,
815
+ "loss": 0.489,
816
+ "mean_token_accuracy": 0.8549116723561353,
817
+ "step": 505
818
+ },
819
+ {
820
+ "epoch": 0.5935408786732616,
821
+ "grad_norm": 0.3960190713405609,
822
+ "learning_rate": 8.481834642606287e-06,
823
+ "loss": 0.4804,
824
+ "mean_token_accuracy": 0.8588179330924269,
825
+ "step": 510
826
+ },
827
+ {
828
+ "epoch": 0.5993599068955484,
829
+ "grad_norm": 0.4174754321575165,
830
+ "learning_rate": 8.281309546109649e-06,
831
+ "loss": 0.4904,
832
+ "mean_token_accuracy": 0.8553032555444215,
833
+ "step": 515
834
+ },
835
+ {
836
+ "epoch": 0.6051789351178353,
837
+ "grad_norm": 0.41682904958724976,
838
+ "learning_rate": 8.081494131116588e-06,
839
+ "loss": 0.4797,
840
+ "mean_token_accuracy": 0.85778211568845,
841
+ "step": 520
842
+ },
843
+ {
844
+ "epoch": 0.6109979633401222,
845
+ "grad_norm": 0.39849600195884705,
846
+ "learning_rate": 7.882470905389827e-06,
847
+ "loss": 0.4983,
848
+ "mean_token_accuracy": 0.8528337732738089,
849
+ "step": 525
850
+ },
851
+ {
852
+ "epoch": 0.6168169915624091,
853
+ "grad_norm": 0.4062061309814453,
854
+ "learning_rate": 7.684322049581359e-06,
855
+ "loss": 0.4918,
856
+ "mean_token_accuracy": 0.8539959811143045,
857
+ "step": 530
858
+ },
859
+ {
860
+ "epoch": 0.622636019784696,
861
+ "grad_norm": 0.38275521993637085,
862
+ "learning_rate": 7.487129383298433e-06,
863
+ "loss": 0.4741,
864
+ "mean_token_accuracy": 0.8592997100935363,
865
+ "step": 535
866
+ },
867
+ {
868
+ "epoch": 0.6284550480069828,
869
+ "grad_norm": 0.3841171860694885,
870
+ "learning_rate": 7.29097433131861e-06,
871
+ "loss": 0.4721,
872
+ "mean_token_accuracy": 0.859289038573644,
873
+ "step": 540
874
+ },
875
+ {
876
+ "epoch": 0.6342740762292697,
877
+ "grad_norm": 0.41029033064842224,
878
+ "learning_rate": 7.095937889967854e-06,
879
+ "loss": 0.4688,
880
+ "mean_token_accuracy": 0.8600678009977978,
881
+ "step": 545
882
+ },
883
+ {
884
+ "epoch": 0.6400931044515565,
885
+ "grad_norm": 0.4171801209449768,
886
+ "learning_rate": 6.90210059367552e-06,
887
+ "loss": 0.4811,
888
+ "mean_token_accuracy": 0.8564380862600244,
889
+ "step": 550
890
+ },
891
+ {
892
+ "epoch": 0.6459121326738435,
893
+ "grad_norm": 0.41676321625709534,
894
+ "learning_rate": 6.7095424817201035e-06,
895
+ "loss": 0.4775,
896
+ "mean_token_accuracy": 0.8579857639838829,
897
+ "step": 555
898
+ },
899
+ {
900
+ "epoch": 0.6517311608961304,
901
+ "grad_norm": 0.40989112854003906,
902
+ "learning_rate": 6.518343065179414e-06,
903
+ "loss": 0.4916,
904
+ "mean_token_accuracy": 0.8554774335695117,
905
+ "step": 560
906
+ },
907
+ {
908
+ "epoch": 0.6575501891184172,
909
+ "grad_norm": 0.41312697529792786,
910
+ "learning_rate": 6.328581294098887e-06,
911
+ "loss": 0.4841,
912
+ "mean_token_accuracy": 0.8565219937605406,
913
+ "step": 565
914
+ },
915
+ {
916
+ "epoch": 0.6633692173407041,
917
+ "grad_norm": 0.40967249870300293,
918
+ "learning_rate": 6.140335524891518e-06,
919
+ "loss": 0.4806,
920
+ "mean_token_accuracy": 0.8576536743376204,
921
+ "step": 570
922
+ },
923
+ {
924
+ "epoch": 0.669188245562991,
925
+ "grad_norm": 0.39644986391067505,
926
+ "learning_rate": 5.953683487982958e-06,
927
+ "loss": 0.4846,
928
+ "mean_token_accuracy": 0.8560946241771303,
929
+ "step": 575
930
+ },
931
+ {
932
+ "epoch": 0.6750072737852778,
933
+ "grad_norm": 0.40402865409851074,
934
+ "learning_rate": 5.768702255715053e-06,
935
+ "loss": 0.4871,
936
+ "mean_token_accuracy": 0.8555429979464371,
937
+ "step": 580
938
+ },
939
+ {
940
+ "epoch": 0.6808263020075648,
941
+ "grad_norm": 0.4249478578567505,
942
+ "learning_rate": 5.58546821052116e-06,
943
+ "loss": 0.4776,
944
+ "mean_token_accuracy": 0.8574913247699447,
945
+ "step": 585
946
+ },
947
+ {
948
+ "epoch": 0.6866453302298516,
949
+ "grad_norm": 0.41155800223350525,
950
+ "learning_rate": 5.404057013386306e-06,
951
+ "loss": 0.4686,
952
+ "mean_token_accuracy": 0.8604994771907524,
953
+ "step": 590
954
+ },
955
+ {
956
+ "epoch": 0.6924643584521385,
957
+ "grad_norm": 0.4111210107803345,
958
+ "learning_rate": 5.224543572605272e-06,
959
+ "loss": 0.4888,
960
+ "mean_token_accuracy": 0.8553543295438057,
961
+ "step": 595
962
+ },
963
+ {
964
+ "epoch": 0.6982833866744254,
965
+ "grad_norm": 0.40666860342025757,
966
+ "learning_rate": 5.047002012851447e-06,
967
+ "loss": 0.4915,
968
+ "mean_token_accuracy": 0.8555927078584643,
969
+ "step": 600
970
+ },
971
+ {
972
+ "epoch": 0.7041024148967122,
973
+ "grad_norm": 0.41017675399780273,
974
+ "learning_rate": 4.871505644569303e-06,
975
+ "loss": 0.4837,
976
+ "mean_token_accuracy": 0.8566939587596106,
977
+ "step": 605
978
+ },
979
+ {
980
+ "epoch": 0.7099214431189991,
981
+ "grad_norm": 0.4147651195526123,
982
+ "learning_rate": 4.698126933703031e-06,
983
+ "loss": 0.483,
984
+ "mean_token_accuracy": 0.8570666700504006,
985
+ "step": 610
986
+ },
987
+ {
988
+ "epoch": 0.715740471341286,
989
+ "grad_norm": 0.40449991822242737,
990
+ "learning_rate": 4.526937471773919e-06,
991
+ "loss": 0.4689,
992
+ "mean_token_accuracy": 0.860008777707927,
993
+ "step": 615
994
+ },
995
+ {
996
+ "epoch": 0.7215594995635729,
997
+ "grad_norm": 0.40807604789733887,
998
+ "learning_rate": 4.35800794631879e-06,
999
+ "loss": 0.4826,
1000
+ "mean_token_accuracy": 0.8565822278909835,
1001
+ "step": 620
1002
+ },
1003
+ {
1004
+ "epoch": 0.7273785277858598,
1005
+ "grad_norm": 0.39830535650253296,
1006
+ "learning_rate": 4.191408111701693e-06,
1007
+ "loss": 0.4731,
1008
+ "mean_token_accuracy": 0.8593161088649728,
1009
+ "step": 625
1010
+ },
1011
+ {
1012
+ "epoch": 0.7331975560081466,
1013
+ "grad_norm": 0.3948498070240021,
1014
+ "learning_rate": 4.0272067603109646e-06,
1015
+ "loss": 0.4676,
1016
+ "mean_token_accuracy": 0.8615339302083884,
1017
+ "step": 630
1018
+ },
1019
+ {
1020
+ "epoch": 0.7390165842304335,
1021
+ "grad_norm": 0.3837417960166931,
1022
+ "learning_rate": 3.865471694153465e-06,
1023
+ "loss": 0.4638,
1024
+ "mean_token_accuracy": 0.8620474819313568,
1025
+ "step": 635
1026
+ },
1027
+ {
1028
+ "epoch": 0.7448356124527203,
1029
+ "grad_norm": 0.4020540118217468,
1030
+ "learning_rate": 3.706269696857785e-06,
1031
+ "loss": 0.4794,
1032
+ "mean_token_accuracy": 0.8580428959872017,
1033
+ "step": 640
1034
+ },
1035
+ {
1036
+ "epoch": 0.7506546406750073,
1037
+ "grad_norm": 0.3988768756389618,
1038
+ "learning_rate": 3.5496665060979563e-06,
1039
+ "loss": 0.4791,
1040
+ "mean_token_accuracy": 0.857951603270609,
1041
+ "step": 645
1042
+ },
1043
+ {
1044
+ "epoch": 0.7564736688972942,
1045
+ "grad_norm": 0.38841554522514343,
1046
+ "learning_rate": 3.395726786449044e-06,
1047
+ "loss": 0.4795,
1048
+ "mean_token_accuracy": 0.8580783028985468,
1049
+ "step": 650
1050
+ },
1051
+ {
1052
+ "epoch": 0.762292697119581,
1053
+ "grad_norm": 0.3969273865222931,
1054
+ "learning_rate": 3.244514102685881e-06,
1055
+ "loss": 0.4752,
1056
+ "mean_token_accuracy": 0.8586160745112773,
1057
+ "step": 655
1058
+ },
1059
+ {
1060
+ "epoch": 0.7681117253418679,
1061
+ "grad_norm": 0.40074124932289124,
1062
+ "learning_rate": 3.0960908935358904e-06,
1063
+ "loss": 0.4881,
1064
+ "mean_token_accuracy": 0.855172939595195,
1065
+ "step": 660
1066
+ },
1067
+ {
1068
+ "epoch": 0.7739307535641547,
1069
+ "grad_norm": 0.4185120463371277,
1070
+ "learning_rate": 2.9505184458968925e-06,
1071
+ "loss": 0.4689,
1072
+ "mean_token_accuracy": 0.8608836993671453,
1073
+ "step": 665
1074
+ },
1075
+ {
1076
+ "epoch": 0.7797497817864416,
1077
+ "grad_norm": 0.42662736773490906,
1078
+ "learning_rate": 2.807856869530534e-06,
1079
+ "loss": 0.4906,
1080
+ "mean_token_accuracy": 0.854303324792995,
1081
+ "step": 670
1082
+ },
1083
+ {
1084
+ "epoch": 0.7855688100087286,
1085
+ "grad_norm": 0.3838253915309906,
1086
+ "learning_rate": 2.6681650722417517e-06,
1087
+ "loss": 0.4739,
1088
+ "mean_token_accuracy": 0.859679923783397,
1089
+ "step": 675
1090
+ },
1091
+ {
1092
+ "epoch": 0.7913878382310154,
1093
+ "grad_norm": 0.4014156460762024,
1094
+ "learning_rate": 2.5315007355545983e-06,
1095
+ "loss": 0.4647,
1096
+ "mean_token_accuracy": 0.8613118683252459,
1097
+ "step": 680
1098
+ },
1099
+ {
1100
+ "epoch": 0.7972068664533023,
1101
+ "grad_norm": 0.41356369853019714,
1102
+ "learning_rate": 2.3979202908943576e-06,
1103
+ "loss": 0.4647,
1104
+ "mean_token_accuracy": 0.8605610072930571,
1105
+ "step": 685
1106
+ },
1107
+ {
1108
+ "epoch": 0.8030258946755892,
1109
+ "grad_norm": 0.4056306481361389,
1110
+ "learning_rate": 2.267478896285913e-06,
1111
+ "loss": 0.48,
1112
+ "mean_token_accuracy": 0.8584646023840529,
1113
+ "step": 690
1114
+ },
1115
+ {
1116
+ "epoch": 0.808844922897876,
1117
+ "grad_norm": 0.3952696919441223,
1118
+ "learning_rate": 2.140230413577882e-06,
1119
+ "loss": 0.4773,
1120
+ "mean_token_accuracy": 0.8575413077069948,
1121
+ "step": 695
1122
+ },
1123
+ {
1124
+ "epoch": 0.814663951120163,
1125
+ "grad_norm": 0.40555790066719055,
1126
+ "learning_rate": 2.0162273862019965e-06,
1127
+ "loss": 0.4822,
1128
+ "mean_token_accuracy": 0.8564873526871629,
1129
+ "step": 700
1130
+ },
1131
+ {
1132
+ "epoch": 0.8204829793424498,
1133
+ "grad_norm": 0.39178261160850525,
1134
+ "learning_rate": 1.8955210174768857e-06,
1135
+ "loss": 0.4665,
1136
+ "mean_token_accuracy": 0.8613343347514766,
1137
+ "step": 705
1138
+ },
1139
+ {
1140
+ "epoch": 0.8263020075647367,
1141
+ "grad_norm": 0.39173874258995056,
1142
+ "learning_rate": 1.7781611494651729e-06,
1143
+ "loss": 0.4736,
1144
+ "mean_token_accuracy": 0.8586594796262844,
1145
+ "step": 710
1146
+ },
1147
+ {
1148
+ "epoch": 0.8321210357870236,
1149
+ "grad_norm": 0.3980768620967865,
1150
+ "learning_rate": 1.6641962423927294e-06,
1151
+ "loss": 0.4742,
1152
+ "mean_token_accuracy": 0.8592263804363405,
1153
+ "step": 715
1154
+ },
1155
+ {
1156
+ "epoch": 0.8379400640093104,
1157
+ "grad_norm": 0.3935108780860901,
1158
+ "learning_rate": 1.5536733546384574e-06,
1159
+ "loss": 0.4861,
1160
+ "mean_token_accuracy": 0.8557533487541985,
1161
+ "step": 720
1162
+ },
1163
+ {
1164
+ "epoch": 0.8437590922315973,
1165
+ "grad_norm": 0.41904276609420776,
1166
+ "learning_rate": 1.4466381233029781e-06,
1167
+ "loss": 0.4806,
1168
+ "mean_token_accuracy": 0.8569492588154313,
1169
+ "step": 725
1170
+ },
1171
+ {
1172
+ "epoch": 0.8495781204538843,
1173
+ "grad_norm": 0.3956993818283081,
1174
+ "learning_rate": 1.3431347453641253e-06,
1175
+ "loss": 0.4772,
1176
+ "mean_token_accuracy": 0.8591885796899528,
1177
+ "step": 730
1178
+ },
1179
+ {
1180
+ "epoch": 0.8553971486761711,
1181
+ "grad_norm": 0.413522869348526,
1182
+ "learning_rate": 1.2432059594271684e-06,
1183
+ "loss": 0.4737,
1184
+ "mean_token_accuracy": 0.8588345902893689,
1185
+ "step": 735
1186
+ },
1187
+ {
1188
+ "epoch": 0.861216176898458,
1189
+ "grad_norm": 0.40007099509239197,
1190
+ "learning_rate": 1.1468930280771728e-06,
1191
+ "loss": 0.4713,
1192
+ "mean_token_accuracy": 0.8596527738293955,
1193
+ "step": 740
1194
+ },
1195
+ {
1196
+ "epoch": 0.8670352051207448,
1197
+ "grad_norm": 0.4135393798351288,
1198
+ "learning_rate": 1.0542357208408572e-06,
1199
+ "loss": 0.5017,
1200
+ "mean_token_accuracy": 0.8527636202781883,
1201
+ "step": 745
1202
+ },
1203
+ {
1204
+ "epoch": 0.8728542333430317,
1205
+ "grad_norm": 0.404310941696167,
1206
+ "learning_rate": 9.652722977649797e-07,
1207
+ "loss": 0.453,
1208
+ "mean_token_accuracy": 0.8641921613151137,
1209
+ "step": 750
1210
+ },
1211
+ {
1212
+ "epoch": 0.8786732615653186,
1213
+ "grad_norm": 0.3922938406467438,
1214
+ "learning_rate": 8.800394936179812e-07,
1215
+ "loss": 0.4654,
1216
+ "mean_token_accuracy": 0.8607802659264431,
1217
+ "step": 755
1218
+ },
1219
+ {
1220
+ "epoch": 0.8844922897876055,
1221
+ "grad_norm": 0.39852988719940186,
1222
+ "learning_rate": 7.985725027214841e-07,
1223
+ "loss": 0.4657,
1224
+ "mean_token_accuracy": 0.8610265562422169,
1225
+ "step": 760
1226
+ },
1227
+ {
1228
+ "epoch": 0.8903113180098924,
1229
+ "grad_norm": 0.4099702835083008,
1230
+ "learning_rate": 7.209049644178489e-07,
1231
+ "loss": 0.4645,
1232
+ "mean_token_accuracy": 0.8611578725473782,
1233
+ "step": 765
1234
+ },
1235
+ {
1236
+ "epoch": 0.8961303462321792,
1237
+ "grad_norm": 0.41344019770622253,
1238
+ "learning_rate": 6.470689491798232e-07,
1239
+ "loss": 0.4709,
1240
+ "mean_token_accuracy": 0.8593041269818004,
1241
+ "step": 770
1242
+ },
1243
+ {
1244
+ "epoch": 0.9019493744544661,
1245
+ "grad_norm": 0.38776156306266785,
1246
+ "learning_rate": 5.770949453680064e-07,
1247
+ "loss": 0.4719,
1248
+ "mean_token_accuracy": 0.8599895481770243,
1249
+ "step": 775
1250
+ },
1251
+ {
1252
+ "epoch": 0.907768402676753,
1253
+ "grad_norm": 0.39070042967796326,
1254
+ "learning_rate": 5.110118466415903e-07,
1255
+ "loss": 0.485,
1256
+ "mean_token_accuracy": 0.8541616461336453,
1257
+ "step": 780
1258
+ },
1259
+ {
1260
+ "epoch": 0.9135874308990398,
1261
+ "grad_norm": 0.39178237318992615,
1262
+ "learning_rate": 4.4884694002760297e-07,
1263
+ "loss": 0.4733,
1264
+ "mean_token_accuracy": 0.8594136677203281,
1265
+ "step": 785
1266
+ },
1267
+ {
1268
+ "epoch": 0.9194064591213268,
1269
+ "grad_norm": 0.3782612383365631,
1270
+ "learning_rate": 3.906258946535446e-07,
1271
+ "loss": 0.4762,
1272
+ "mean_token_accuracy": 0.8573084054617641,
1273
+ "step": 790
1274
+ },
1275
+ {
1276
+ "epoch": 0.9252254873436137,
1277
+ "grad_norm": 0.39697375893592834,
1278
+ "learning_rate": 3.363727511480974e-07,
1279
+ "loss": 0.4765,
1280
+ "mean_token_accuracy": 0.8577650392622143,
1281
+ "step": 795
1282
+ },
1283
+ {
1284
+ "epoch": 0.9310445155659005,
1285
+ "grad_norm": 0.3849934935569763,
1286
+ "learning_rate": 2.861099117142718e-07,
1287
+ "loss": 0.461,
1288
+ "mean_token_accuracy": 0.8616937448944618,
1289
+ "step": 800
1290
+ },
1291
+ {
1292
+ "epoch": 0.9368635437881874,
1293
+ "grad_norm": 0.4121275842189789,
1294
+ "learning_rate": 2.398581308790848e-07,
1295
+ "loss": 0.4827,
1296
+ "mean_token_accuracy": 0.857346642587542,
1297
+ "step": 805
1298
+ },
1299
+ {
1300
+ "epoch": 0.9426825720104742,
1301
+ "grad_norm": 0.39553239941596985,
1302
+ "learning_rate": 1.976365069236108e-07,
1303
+ "loss": 0.475,
1304
+ "mean_token_accuracy": 0.8578932364447682,
1305
+ "step": 810
1306
+ },
1307
+ {
1308
+ "epoch": 0.9485016002327611,
1309
+ "grad_norm": 0.3881802558898926,
1310
+ "learning_rate": 1.5946247399691638e-07,
1311
+ "loss": 0.4706,
1312
+ "mean_token_accuracy": 0.8601075021722997,
1313
+ "step": 815
1314
+ },
1315
+ {
1316
+ "epoch": 0.9543206284550481,
1317
+ "grad_norm": 0.3950815498828888,
1318
+ "learning_rate": 1.2535179491715453e-07,
1319
+ "loss": 0.4666,
1320
+ "mean_token_accuracy": 0.8601555329046816,
1321
+ "step": 820
1322
+ },
1323
+ {
1324
+ "epoch": 0.9601396566773349,
1325
+ "grad_norm": 0.4037560522556305,
1326
+ "learning_rate": 9.531855466278218e-08,
1327
+ "loss": 0.4636,
1328
+ "mean_token_accuracy": 0.8610238882101969,
1329
+ "step": 825
1330
+ },
1331
+ {
1332
+ "epoch": 0.9659586848996218,
1333
+ "grad_norm": 0.3907164931297302,
1334
+ "learning_rate": 6.937515455659128e-08,
1335
+ "loss": 0.4686,
1336
+ "mean_token_accuracy": 0.8608224325086367,
1337
+ "step": 830
1338
+ },
1339
+ {
1340
+ "epoch": 0.9717777131219086,
1341
+ "grad_norm": 0.41244828701019287,
1342
+ "learning_rate": 4.753230714496071e-08,
1343
+ "loss": 0.4832,
1344
+ "mean_token_accuracy": 0.8551789295502588,
1345
+ "step": 835
1346
+ },
1347
+ {
1348
+ "epoch": 0.9775967413441955,
1349
+ "grad_norm": 0.39467689394950867,
1350
+ "learning_rate": 2.9799031774427888e-08,
1351
+ "loss": 0.4877,
1352
+ "mean_token_accuracy": 0.8558349555197562,
1353
+ "step": 840
1354
+ },
1355
+ {
1356
+ "epoch": 0.9834157695664824,
1357
+ "grad_norm": 0.39321061968803406,
1358
+ "learning_rate": 1.6182650867421213e-08,
1359
+ "loss": 0.4634,
1360
+ "mean_token_accuracy": 0.8616077323763098,
1361
+ "step": 845
1362
+ },
1363
+ {
1364
+ "epoch": 0.9892347977887693,
1365
+ "grad_norm": 0.4110186696052551,
1366
+ "learning_rate": 6.688786898688682e-09,
1367
+ "loss": 0.4885,
1368
+ "mean_token_accuracy": 0.8555144290494805,
1369
+ "step": 850
1370
+ },
1371
+ {
1372
+ "epoch": 0.9950538260110562,
1373
+ "grad_norm": 0.388217955827713,
1374
+ "learning_rate": 1.321360073662792e-09,
1375
+ "loss": 0.4607,
1376
+ "mean_token_accuracy": 0.8616522037739971,
1377
+ "step": 855
1378
+ },
1379
+ {
1380
+ "epoch": 0.9997090485888857,
1381
+ "mean_token_accuracy": 0.8550544311901966,
1382
+ "step": 859,
1383
+ "total_flos": 42703635087360.0,
1384
+ "train_loss": 0.504338482067388,
1385
+ "train_runtime": 61692.734,
1386
+ "train_samples_per_second": 0.446,
1387
+ "train_steps_per_second": 0.014
1388
+ }
1389
+ ],
1390
+ "logging_steps": 5,
1391
+ "max_steps": 859,
1392
+ "num_input_tokens_seen": 0,
1393
+ "num_train_epochs": 1,
1394
+ "save_steps": 500,
1395
+ "stateful_callbacks": {
1396
+ "TrainerControl": {
1397
+ "args": {
1398
+ "should_epoch_stop": false,
1399
+ "should_evaluate": false,
1400
+ "should_log": false,
1401
+ "should_save": false,
1402
+ "should_training_stop": false
1403
+ },
1404
+ "attributes": {}
1405
+ }
1406
+ },
1407
+ "total_flos": 42703635087360.0,
1408
+ "train_batch_size": 2,
1409
+ "trial_name": null,
1410
+ "trial_params": null
1411
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:983798bab29c360829b545be7673fc80123f0e34e818caa4b695ab09cfd444ff
3
+ size 7416