Delete folder llama_factory with huggingface_hub

llama_factory/v2-20241220-pos/README.md

@@ -1,62 +0,0 @@
----
-library_name: peft
-license: other
-base_model: /disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct
-tags:
-- llama-factory
-- lora
-- generated_from_trainer
-model-index:
-- name: v2-20241220-pos
-  results: []
----
-
-<!-- This model card has been generated automatically according to the information the Trainer had access to. You
-should probably proofread and complete it, then remove this comment. -->
-
-# v2-20241220-pos
-
-This model is a fine-tuned version of [/disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct](https://huggingface.co//disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct) on the M^2RAR-20241220-LLM-Pos dataset.
-
-## Model description
-
-More information needed
-
-## Intended uses & limitations
-
-More information needed
-
-## Training and evaluation data
-
-More information needed
-
-## Training procedure
-
-### Training hyperparameters
-
-The following hyperparameters were used during training:
-- learning_rate: 0.0001
-- train_batch_size: 2
-- eval_batch_size: 8
-- seed: 42
-- distributed_type: multi-GPU
-- num_devices: 2
-- gradient_accumulation_steps: 16
-- total_train_batch_size: 64
-- total_eval_batch_size: 16
-- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
-- lr_scheduler_type: cosine
-- lr_scheduler_warmup_ratio: 0.1
-- num_epochs: 3.0
-
-### Training results
-
-
-
-### Framework versions
-
-- PEFT 0.12.0
-- Transformers 4.46.1
-- Pytorch 2.3.1+cu121
-- Datasets 3.1.0
-- Tokenizers 0.20.3

llama_factory/v2-20241220-pos/adapter_config.json

@@ -1,34 +0,0 @@
-{
-  "alpha_pattern": {},
-  "auto_mapping": null,
-  "base_model_name_or_path": "/disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct",
-  "bias": "none",
-  "fan_in_fan_out": false,
-  "inference_mode": true,
-  "init_lora_weights": true,
-  "layer_replication": null,
-  "layers_pattern": null,
-  "layers_to_transform": null,
-  "loftq_config": {},
-  "lora_alpha": 256,
-  "lora_dropout": 0.0,
-  "megatron_config": null,
-  "megatron_core": "megatron.core",
-  "modules_to_save": null,
-  "peft_type": "LORA",
-  "r": 128,
-  "rank_pattern": {},
-  "revision": null,
-  "target_modules": [
-    "q_proj",
-    "v_proj",
-    "down_proj",
-    "up_proj",
-    "k_proj",
-    "gate_proj",
-    "o_proj"
-  ],
-  "task_type": "CAUSAL_LM",
-  "use_dora": false,
-  "use_rslora": false
-}

llama_factory/v2-20241220-pos/added_tokens.json

@@ -1,24 +0,0 @@
-{
-  "</tool_call>": 151658,
-  "<tool_call>": 151657,
-  "<|box_end|>": 151649,
-  "<|box_start|>": 151648,
-  "<|endoftext|>": 151643,
-  "<|file_sep|>": 151664,
-  "<|fim_middle|>": 151660,
-  "<|fim_pad|>": 151662,
-  "<|fim_prefix|>": 151659,
-  "<|fim_suffix|>": 151661,
-  "<|im_end|>": 151645,
-  "<|im_start|>": 151644,
-  "<|image_pad|>": 151655,
-  "<|object_ref_end|>": 151647,
-  "<|object_ref_start|>": 151646,
-  "<|quad_end|>": 151651,
-  "<|quad_start|>": 151650,
-  "<|repo_name|>": 151663,
-  "<|video_pad|>": 151656,
-  "<|vision_end|>": 151653,
-  "<|vision_pad|>": 151654,
-  "<|vision_start|>": 151652
-}

llama_factory/v2-20241220-pos/all_results.json

@@ -1,8 +0,0 @@
-{
-    "epoch": 2.56,
-    "total_flos": 3.465249321732014e+17,
-    "train_loss": 1.5466326872507732,
-    "train_runtime": 1519.4813,
-    "train_samples_per_second": 0.592,
-    "train_steps_per_second": 0.008
-}

llama_factory/v2-20241220-pos/special_tokens_map.json

@@ -1,31 +0,0 @@
-{
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>",
-    "<|object_ref_start|>",
-    "<|object_ref_end|>",
-    "<|box_start|>",
-    "<|box_end|>",
-    "<|quad_start|>",
-    "<|quad_end|>",
-    "<|vision_start|>",
-    "<|vision_end|>",
-    "<|vision_pad|>",
-    "<|image_pad|>",
-    "<|video_pad|>"
-  ],
-  "eos_token": {
-    "content": "<|im_end|>",
-    "lstrip": false,
-    "normalized": false,
-    "rstrip": false,
-    "single_word": false
-  },
-  "pad_token": {
-    "content": "<|endoftext|>",
-    "lstrip": false,
-    "normalized": false,
-    "rstrip": false,
-    "single_word": false
-  }
-}

llama_factory/v2-20241220-pos/tokenizer_config.json

@@ -1,208 +0,0 @@
-{
-  "add_bos_token": false,
-  "add_prefix_space": false,
-  "added_tokens_decoder": {
-    "151643": {
-      "content": "<|endoftext|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151644": {
-      "content": "<|im_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151645": {
-      "content": "<|im_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151646": {
-      "content": "<|object_ref_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151647": {
-      "content": "<|object_ref_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151648": {
-      "content": "<|box_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151649": {
-      "content": "<|box_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151650": {
-      "content": "<|quad_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151651": {
-      "content": "<|quad_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151652": {
-      "content": "<|vision_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151653": {
-      "content": "<|vision_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151654": {
-      "content": "<|vision_pad|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151655": {
-      "content": "<|image_pad|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151656": {
-      "content": "<|video_pad|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151657": {
-      "content": "<tool_call>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151658": {
-      "content": "</tool_call>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151659": {
-      "content": "<|fim_prefix|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151660": {
-      "content": "<|fim_middle|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151661": {
-      "content": "<|fim_suffix|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151662": {
-      "content": "<|fim_pad|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151663": {
-      "content": "<|repo_name|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151664": {
-      "content": "<|file_sep|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    }
-  },
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>",
-    "<|object_ref_start|>",
-    "<|object_ref_end|>",
-    "<|box_start|>",
-    "<|box_end|>",
-    "<|quad_start|>",
-    "<|quad_end|>",
-    "<|vision_start|>",
-    "<|vision_end|>",
-    "<|vision_pad|>",
-    "<|image_pad|>",
-    "<|video_pad|>"
-  ],
-  "bos_token": null,
-  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
-  "clean_up_tokenization_spaces": false,
-  "eos_token": "<|im_end|>",
-  "errors": "replace",
-  "model_max_length": 131072,
-  "pad_token": "<|endoftext|>",
-  "padding_side": "right",
-  "split_special_tokens": false,
-  "tokenizer_class": "Qwen2Tokenizer",
-  "unk_token": null
-}

llama_factory/v2-20241220-pos/train_results.json

@@ -1,8 +0,0 @@
-{
-    "epoch": 2.56,
-    "total_flos": 3.465249321732014e+17,
-    "train_loss": 1.5466326872507732,
-    "train_runtime": 1519.4813,
-    "train_samples_per_second": 0.592,
-    "train_steps_per_second": 0.008
-}

llama_factory/v2-20241220-pos/trainer_log.jsonl

@@ -1,13 +0,0 @@
-{"current_steps": 1, "total_steps": 12, "loss": 1.6841, "lr": 5e-05, "epoch": 0.21333333333333335, "percentage": 8.33, "elapsed_time": "0:02:04", "remaining_time": "0:22:48"}
-{"current_steps": 2, "total_steps": 12, "loss": 1.6529, "lr": 0.0001, "epoch": 0.4266666666666667, "percentage": 16.67, "elapsed_time": "0:04:27", "remaining_time": "0:22:15"}
-{"current_steps": 3, "total_steps": 12, "loss": 1.5033, "lr": 9.755282581475769e-05, "epoch": 0.64, "percentage": 25.0, "elapsed_time": "0:06:27", "remaining_time": "0:19:23"}
-{"current_steps": 4, "total_steps": 12, "loss": 1.5099, "lr": 9.045084971874738e-05, "epoch": 0.8533333333333334, "percentage": 33.33, "elapsed_time": "0:08:41", "remaining_time": "0:17:23"}
-{"current_steps": 5, "total_steps": 12, "loss": 1.9194, "lr": 7.938926261462366e-05, "epoch": 1.0666666666666667, "percentage": 41.67, "elapsed_time": "0:10:24", "remaining_time": "0:14:33"}
-{"current_steps": 6, "total_steps": 12, "loss": 1.4928, "lr": 6.545084971874738e-05, "epoch": 1.28, "percentage": 50.0, "elapsed_time": "0:12:24", "remaining_time": "0:12:24"}
-{"current_steps": 7, "total_steps": 12, "loss": 1.3846, "lr": 5e-05, "epoch": 1.4933333333333334, "percentage": 58.33, "elapsed_time": "0:14:22", "remaining_time": "0:10:15"}
-{"current_steps": 8, "total_steps": 12, "loss": 1.3936, "lr": 3.4549150281252636e-05, "epoch": 1.7066666666666666, "percentage": 66.67, "elapsed_time": "0:16:13", "remaining_time": "0:08:06"}
-{"current_steps": 9, "total_steps": 12, "loss": 1.5382, "lr": 2.061073738537635e-05, "epoch": 1.92, "percentage": 75.0, "elapsed_time": "0:18:11", "remaining_time": "0:06:03"}
-{"current_steps": 10, "total_steps": 12, "loss": 1.762, "lr": 9.549150281252633e-06, "epoch": 2.1333333333333333, "percentage": 83.33, "elapsed_time": "0:20:47", "remaining_time": "0:04:09"}
-{"current_steps": 11, "total_steps": 12, "loss": 1.2744, "lr": 2.4471741852423237e-06, "epoch": 2.3466666666666667, "percentage": 91.67, "elapsed_time": "0:22:28", "remaining_time": "0:02:02"}
-{"current_steps": 12, "total_steps": 12, "loss": 1.4444, "lr": 0.0, "epoch": 2.56, "percentage": 100.0, "elapsed_time": "0:24:52", "remaining_time": "0:00:00"}
-{"current_steps": 12, "total_steps": 12, "epoch": 2.56, "percentage": 100.0, "elapsed_time": "0:25:19", "remaining_time": "0:00:00"}

llama_factory/v2-20241220-pos/trainer_state.json

@@ -1,126 +0,0 @@
-{
-  "best_metric": null,
-  "best_model_checkpoint": null,
-  "epoch": 2.56,
-  "eval_steps": 500,
-  "global_step": 12,
-  "is_hyper_param_search": false,
-  "is_local_process_zero": true,
-  "is_world_process_zero": true,
-  "log_history": [
-    {
-      "epoch": 0.21333333333333335,
-      "grad_norm": 1.4990226030349731,
-      "learning_rate": 5e-05,
-      "loss": 1.6841,
-      "step": 1
-    },
-    {
-      "epoch": 0.4266666666666667,
-      "grad_norm": 1.4839531183242798,
-      "learning_rate": 0.0001,
-      "loss": 1.6529,
-      "step": 2
-    },
-    {
-      "epoch": 0.64,
-      "grad_norm": 0.4049980044364929,
-      "learning_rate": 9.755282581475769e-05,
-      "loss": 1.5033,
-      "step": 3
-    },
-    {
-      "epoch": 0.8533333333333334,
-      "grad_norm": 0.7275316715240479,
-      "learning_rate": 9.045084971874738e-05,
-      "loss": 1.5099,
-      "step": 4
-    },
-    {
-      "epoch": 1.0666666666666667,
-      "grad_norm": 0.6615236401557922,
-      "learning_rate": 7.938926261462366e-05,
-      "loss": 1.9194,
-      "step": 5
-    },
-    {
-      "epoch": 1.28,
-      "grad_norm": 0.3167978525161743,
-      "learning_rate": 6.545084971874738e-05,
-      "loss": 1.4928,
-      "step": 6
-    },
-    {
-      "epoch": 1.4933333333333334,
-      "grad_norm": 0.22702914476394653,
-      "learning_rate": 5e-05,
-      "loss": 1.3846,
-      "step": 7
-    },
-    {
-      "epoch": 1.7066666666666666,
-      "grad_norm": 0.2197398990392685,
-      "learning_rate": 3.4549150281252636e-05,
-      "loss": 1.3936,
-      "step": 8
-    },
-    {
-      "epoch": 1.92,
-      "grad_norm": 0.2682381272315979,
-      "learning_rate": 2.061073738537635e-05,
-      "loss": 1.5382,
-      "step": 9
-    },
-    {
-      "epoch": 2.1333333333333333,
-      "grad_norm": 0.2999463975429535,
-      "learning_rate": 9.549150281252633e-06,
-      "loss": 1.762,
-      "step": 10
-    },
-    {
-      "epoch": 2.3466666666666667,
-      "grad_norm": 0.1912921667098999,
-      "learning_rate": 2.4471741852423237e-06,
-      "loss": 1.2744,
-      "step": 11
-    },
-    {
-      "epoch": 2.56,
-      "grad_norm": 0.20647935569286346,
-      "learning_rate": 0.0,
-      "loss": 1.4444,
-      "step": 12
-    },
-    {
-      "epoch": 2.56,
-      "step": 12,
-      "total_flos": 3.465249321732014e+17,
-      "train_loss": 1.5466326872507732,
-      "train_runtime": 1519.4813,
-      "train_samples_per_second": 0.592,
-      "train_steps_per_second": 0.008
-    }
-  ],
-  "logging_steps": 1,
-  "max_steps": 12,
-  "num_input_tokens_seen": 0,
-  "num_train_epochs": 3,
-  "save_steps": 200,
-  "stateful_callbacks": {
-    "TrainerControl": {
-      "args": {
-        "should_epoch_stop": false,
-        "should_evaluate": false,
-        "should_log": false,
-        "should_save": true,
-        "should_training_stop": true
-      },
-      "attributes": {}
-    }
-  },
-  "total_flos": 3.465249321732014e+17,
-  "train_batch_size": 2,
-  "trial_name": null,
-  "trial_params": null
-}

llama_factory/v2-20241220/README.md

@@ -1,62 +0,0 @@
----
-library_name: peft
-license: other
-base_model: /disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct
-tags:
-- llama-factory
-- lora
-- generated_from_trainer
-model-index:
-- name: llama_factory
-  results: []
----
-
-<!-- This model card has been generated automatically according to the information the Trainer had access to. You
-should probably proofread and complete it, then remove this comment. -->
-
-# llama_factory
-
-This model is a fine-tuned version of [/disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct](https://huggingface.co//disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct) on the M^2RAR-20241220-LLM dataset.
-
-## Model description
-
-More information needed
-
-## Intended uses & limitations
-
-More information needed
-
-## Training and evaluation data
-
-More information needed
-
-## Training procedure
-
-### Training hyperparameters
-
-The following hyperparameters were used during training:
-- learning_rate: 0.0001
-- train_batch_size: 2
-- eval_batch_size: 8
-- seed: 42
-- distributed_type: multi-GPU
-- num_devices: 2
-- gradient_accumulation_steps: 16
-- total_train_batch_size: 64
-- total_eval_batch_size: 16
-- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
-- lr_scheduler_type: cosine
-- lr_scheduler_warmup_ratio: 0.1
-- num_epochs: 3.0
-
-### Training results
-
-
-
-### Framework versions
-
-- PEFT 0.12.0
-- Transformers 4.46.1
-- Pytorch 2.3.1+cu121
-- Datasets 3.1.0
-- Tokenizers 0.20.3

llama_factory/v2-20241220/adapter_config.json

@@ -1,34 +0,0 @@
-{
-  "alpha_pattern": {},
-  "auto_mapping": null,
-  "base_model_name_or_path": "/disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct",
-  "bias": "none",
-  "fan_in_fan_out": false,
-  "inference_mode": true,
-  "init_lora_weights": true,
-  "layer_replication": null,
-  "layers_pattern": null,
-  "layers_to_transform": null,
-  "loftq_config": {},
-  "lora_alpha": 256,
-  "lora_dropout": 0.0,
-  "megatron_config": null,
-  "megatron_core": "megatron.core",
-  "modules_to_save": null,
-  "peft_type": "LORA",
-  "r": 128,
-  "rank_pattern": {},
-  "revision": null,
-  "target_modules": [
-    "up_proj",
-    "o_proj",
-    "k_proj",
-    "v_proj",
-    "down_proj",
-    "q_proj",
-    "gate_proj"
-  ],
-  "task_type": "CAUSAL_LM",
-  "use_dora": false,
-  "use_rslora": false
-}

llama_factory/v2-20241220/added_tokens.json

@@ -1,24 +0,0 @@
-{
-  "</tool_call>": 151658,
-  "<tool_call>": 151657,
-  "<|box_end|>": 151649,
-  "<|box_start|>": 151648,
-  "<|endoftext|>": 151643,
-  "<|file_sep|>": 151664,
-  "<|fim_middle|>": 151660,
-  "<|fim_pad|>": 151662,
-  "<|fim_prefix|>": 151659,
-  "<|fim_suffix|>": 151661,
-  "<|im_end|>": 151645,
-  "<|im_start|>": 151644,
-  "<|image_pad|>": 151655,
-  "<|object_ref_end|>": 151647,
-  "<|object_ref_start|>": 151646,
-  "<|quad_end|>": 151651,
-  "<|quad_start|>": 151650,
-  "<|repo_name|>": 151663,
-  "<|video_pad|>": 151656,
-  "<|vision_end|>": 151653,
-  "<|vision_pad|>": 151654,
-  "<|vision_start|>": 151652
-}

llama_factory/v2-20241220/all_results.json

@@ -1,8 +0,0 @@
-{
-    "epoch": 2.88,
-    "total_flos": 5.0233808000752026e+17,
-    "train_loss": 1.5231221450699701,
-    "train_runtime": 2241.9981,
-    "train_samples_per_second": 0.535,
-    "train_steps_per_second": 0.008
-}

llama_factory/v2-20241220/special_tokens_map.json

@@ -1,31 +0,0 @@
-{
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>",
-    "<|object_ref_start|>",
-    "<|object_ref_end|>",
-    "<|box_start|>",
-    "<|box_end|>",
-    "<|quad_start|>",
-    "<|quad_end|>",
-    "<|vision_start|>",
-    "<|vision_end|>",
-    "<|vision_pad|>",
-    "<|image_pad|>",
-    "<|video_pad|>"
-  ],
-  "eos_token": {
-    "content": "<|im_end|>",
-    "lstrip": false,
-    "normalized": false,
-    "rstrip": false,
-    "single_word": false
-  },
-  "pad_token": {
-    "content": "<|endoftext|>",
-    "lstrip": false,
-    "normalized": false,
-    "rstrip": false,
-    "single_word": false
-  }
-}

llama_factory/v2-20241220/tokenizer_config.json

@@ -1,208 +0,0 @@
-{
-  "add_bos_token": false,
-  "add_prefix_space": false,
-  "added_tokens_decoder": {
-    "151643": {
-      "content": "<|endoftext|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151644": {
-      "content": "<|im_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151645": {
-      "content": "<|im_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151646": {
-      "content": "<|object_ref_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151647": {
-      "content": "<|object_ref_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151648": {
-      "content": "<|box_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151649": {
-      "content": "<|box_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151650": {
-      "content": "<|quad_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151651": {
-      "content": "<|quad_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151652": {
-      "content": "<|vision_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151653": {
-      "content": "<|vision_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151654": {
-      "content": "<|vision_pad|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151655": {
-      "content": "<|image_pad|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151656": {
-      "content": "<|video_pad|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151657": {
-      "content": "<tool_call>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151658": {
-      "content": "</tool_call>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151659": {
-      "content": "<|fim_prefix|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151660": {
-      "content": "<|fim_middle|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151661": {
-      "content": "<|fim_suffix|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151662": {
-      "content": "<|fim_pad|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151663": {
-      "content": "<|repo_name|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151664": {
-      "content": "<|file_sep|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    }
-  },
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>",
-    "<|object_ref_start|>",
-    "<|object_ref_end|>",
-    "<|box_start|>",
-    "<|box_end|>",
-    "<|quad_start|>",
-    "<|quad_end|>",
-    "<|vision_start|>",
-    "<|vision_end|>",
-    "<|vision_pad|>",
-    "<|image_pad|>",
-    "<|video_pad|>"
-  ],
-  "bos_token": null,
-  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
-  "clean_up_tokenization_spaces": false,
-  "eos_token": "<|im_end|>",
-  "errors": "replace",
-  "model_max_length": 131072,
-  "pad_token": "<|endoftext|>",
-  "padding_side": "right",
-  "split_special_tokens": false,
-  "tokenizer_class": "Qwen2Tokenizer",
-  "unk_token": null
-}

llama_factory/v2-20241220/train_results.json

@@ -1,8 +0,0 @@
-{
-    "epoch": 2.88,
-    "total_flos": 5.0233808000752026e+17,
-    "train_loss": 1.5231221450699701,
-    "train_runtime": 2241.9981,
-    "train_samples_per_second": 0.535,
-    "train_steps_per_second": 0.008
-}

llama_factory/v2-20241220/trainer_log.jsonl

@@ -1,19 +0,0 @@
-{"current_steps": 1, "total_steps": 18, "loss": 1.5935, "lr": 5e-05, "epoch": 0.16, "percentage": 5.56, "elapsed_time": "0:02:16", "remaining_time": "0:38:42"}
-{"current_steps": 2, "total_steps": 18, "loss": 1.5899, "lr": 0.0001, "epoch": 0.32, "percentage": 11.11, "elapsed_time": "0:04:18", "remaining_time": "0:34:26"}
-{"current_steps": 3, "total_steps": 18, "loss": 1.4721, "lr": 9.903926402016153e-05, "epoch": 0.48, "percentage": 16.67, "elapsed_time": "0:06:12", "remaining_time": "0:31:04"}
-{"current_steps": 4, "total_steps": 18, "loss": 1.5331, "lr": 9.619397662556435e-05, "epoch": 0.64, "percentage": 22.22, "elapsed_time": "0:08:08", "remaining_time": "0:28:30"}
-{"current_steps": 5, "total_steps": 18, "loss": 1.4734, "lr": 9.157348061512727e-05, "epoch": 0.8, "percentage": 27.78, "elapsed_time": "0:10:26", "remaining_time": "0:27:09"}
-{"current_steps": 6, "total_steps": 18, "loss": 1.4631, "lr": 8.535533905932738e-05, "epoch": 0.96, "percentage": 33.33, "elapsed_time": "0:12:44", "remaining_time": "0:25:28"}
-{"current_steps": 7, "total_steps": 18, "loss": 2.5345, "lr": 7.777851165098012e-05, "epoch": 1.12, "percentage": 38.89, "elapsed_time": "0:14:44", "remaining_time": "0:23:09"}
-{"current_steps": 8, "total_steps": 18, "loss": 1.4092, "lr": 6.91341716182545e-05, "epoch": 1.28, "percentage": 44.44, "elapsed_time": "0:17:10", "remaining_time": "0:21:27"}
-{"current_steps": 9, "total_steps": 18, "loss": 1.4112, "lr": 5.9754516100806423e-05, "epoch": 1.44, "percentage": 50.0, "elapsed_time": "0:19:05", "remaining_time": "0:19:05"}
-{"current_steps": 10, "total_steps": 18, "loss": 1.3291, "lr": 5e-05, "epoch": 1.6, "percentage": 55.56, "elapsed_time": "0:21:09", "remaining_time": "0:16:55"}
-{"current_steps": 11, "total_steps": 18, "loss": 1.3225, "lr": 4.0245483899193595e-05, "epoch": 1.76, "percentage": 61.11, "elapsed_time": "0:23:09", "remaining_time": "0:14:44"}
-{"current_steps": 12, "total_steps": 18, "loss": 1.3646, "lr": 3.086582838174551e-05, "epoch": 1.92, "percentage": 66.67, "elapsed_time": "0:25:06", "remaining_time": "0:12:33"}
-{"current_steps": 13, "total_steps": 18, "loss": 2.3625, "lr": 2.2221488349019903e-05, "epoch": 2.08, "percentage": 72.22, "elapsed_time": "0:27:09", "remaining_time": "0:10:26"}
-{"current_steps": 14, "total_steps": 18, "loss": 1.3055, "lr": 1.4644660940672627e-05, "epoch": 2.24, "percentage": 77.78, "elapsed_time": "0:28:52", "remaining_time": "0:08:15"}
-{"current_steps": 15, "total_steps": 18, "loss": 1.3382, "lr": 8.426519384872733e-06, "epoch": 2.4, "percentage": 83.33, "elapsed_time": "0:30:40", "remaining_time": "0:06:08"}
-{"current_steps": 16, "total_steps": 18, "loss": 1.3593, "lr": 3.8060233744356633e-06, "epoch": 2.56, "percentage": 88.89, "elapsed_time": "0:32:34", "remaining_time": "0:04:04"}
-{"current_steps": 17, "total_steps": 18, "loss": 1.2852, "lr": 9.607359798384785e-07, "epoch": 2.7199999999999998, "percentage": 94.44, "elapsed_time": "0:34:38", "remaining_time": "0:02:02"}
-{"current_steps": 18, "total_steps": 18, "loss": 1.2693, "lr": 0.0, "epoch": 2.88, "percentage": 100.0, "elapsed_time": "0:36:51", "remaining_time": "0:00:00"}
-{"current_steps": 18, "total_steps": 18, "epoch": 2.88, "percentage": 100.0, "elapsed_time": "0:37:21", "remaining_time": "0:00:00"}

llama_factory/v2-20241220/trainer_state.json

@@ -1,168 +0,0 @@
-{
-  "best_metric": null,
-  "best_model_checkpoint": null,
-  "epoch": 2.88,
-  "eval_steps": 500,
-  "global_step": 18,
-  "is_hyper_param_search": false,
-  "is_local_process_zero": true,
-  "is_world_process_zero": true,
-  "log_history": [
-    {
-      "epoch": 0.16,
-      "grad_norm": 1.4122823476791382,
-      "learning_rate": 5e-05,
-      "loss": 1.5935,
-      "step": 1
-    },
-    {
-      "epoch": 0.32,
-      "grad_norm": 1.3658671379089355,
-      "learning_rate": 0.0001,
-      "loss": 1.5899,
-      "step": 2
-    },
-    {
-      "epoch": 0.48,
-      "grad_norm": 0.41520926356315613,
-      "learning_rate": 9.903926402016153e-05,
-      "loss": 1.4721,
-      "step": 3
-    },
-    {
-      "epoch": 0.64,
-      "grad_norm": 0.7108041048049927,
-      "learning_rate": 9.619397662556435e-05,
-      "loss": 1.5331,
-      "step": 4
-    },
-    {
-      "epoch": 0.8,
-      "grad_norm": 0.5275041460990906,
-      "learning_rate": 9.157348061512727e-05,
-      "loss": 1.4734,
-      "step": 5
-    },
-    {
-      "epoch": 0.96,
-      "grad_norm": 0.3307032883167267,
-      "learning_rate": 8.535533905932738e-05,
-      "loss": 1.4631,
-      "step": 6
-    },
-    {
-      "epoch": 1.12,
-      "grad_norm": 0.44455254077911377,
-      "learning_rate": 7.777851165098012e-05,
-      "loss": 2.5345,
-      "step": 7
-    },
-    {
-      "epoch": 1.28,
-      "grad_norm": 0.21081414818763733,
-      "learning_rate": 6.91341716182545e-05,
-      "loss": 1.4092,
-      "step": 8
-    },
-    {
-      "epoch": 1.44,
-      "grad_norm": 0.261588990688324,
-      "learning_rate": 5.9754516100806423e-05,
-      "loss": 1.4112,
-      "step": 9
-    },
-    {
-      "epoch": 1.6,
-      "grad_norm": 0.23433491587638855,
-      "learning_rate": 5e-05,
-      "loss": 1.3291,
-      "step": 10
-    },
-    {
-      "epoch": 1.76,
-      "grad_norm": 0.1907648742198944,
-      "learning_rate": 4.0245483899193595e-05,
-      "loss": 1.3225,
-      "step": 11
-    },
-    {
-      "epoch": 1.92,
-      "grad_norm": 0.17216548323631287,
-      "learning_rate": 3.086582838174551e-05,
-      "loss": 1.3646,
-      "step": 12
-    },
-    {
-      "epoch": 2.08,
-      "grad_norm": 0.34126514196395874,
-      "learning_rate": 2.2221488349019903e-05,
-      "loss": 2.3625,
-      "step": 13
-    },
-    {
-      "epoch": 2.24,
-      "grad_norm": 0.16684114933013916,
-      "learning_rate": 1.4644660940672627e-05,
-      "loss": 1.3055,
-      "step": 14
-    },
-    {
-      "epoch": 2.4,
-      "grad_norm": 0.17330783605575562,
-      "learning_rate": 8.426519384872733e-06,
-      "loss": 1.3382,
-      "step": 15
-    },
-    {
-      "epoch": 2.56,
-      "grad_norm": 0.1761721968650818,
-      "learning_rate": 3.8060233744356633e-06,
-      "loss": 1.3593,
-      "step": 16
-    },
-    {
-      "epoch": 2.7199999999999998,
-      "grad_norm": 0.16674228012561798,
-      "learning_rate": 9.607359798384785e-07,
-      "loss": 1.2852,
-      "step": 17
-    },
-    {
-      "epoch": 2.88,
-      "grad_norm": 0.16730456054210663,
-      "learning_rate": 0.0,
-      "loss": 1.2693,
-      "step": 18
-    },
-    {
-      "epoch": 2.88,
-      "step": 18,
-      "total_flos": 5.0233808000752026e+17,
-      "train_loss": 1.5231221450699701,
-      "train_runtime": 2241.9981,
-      "train_samples_per_second": 0.535,
-      "train_steps_per_second": 0.008
-    }
-  ],
-  "logging_steps": 1,
-  "max_steps": 18,
-  "num_input_tokens_seen": 0,
-  "num_train_epochs": 3,
-  "save_steps": 200,
-  "stateful_callbacks": {
-    "TrainerControl": {
-      "args": {
-        "should_epoch_stop": false,
-        "should_evaluate": false,
-        "should_log": false,
-        "should_save": true,
-        "should_training_stop": true
-      },
-      "attributes": {}
-    }
-  },
-  "total_flos": 5.0233808000752026e+17,
-  "train_batch_size": 2,
-  "trial_name": null,
-  "trial_params": null
-}

llama_factory/v3-20241225-1_10/README.md

@@ -1,62 +0,0 @@
----
-library_name: peft
-license: other
-base_model: /disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct
-tags:
-- llama-factory
-- lora
-- generated_from_trainer
-model-index:
-- name: v3-20241225-1_10
-  results: []
----
-
-<!-- This model card has been generated automatically according to the information the Trainer had access to. You
-should probably proofread and complete it, then remove this comment. -->
-
-# v3-20241225-1_10
-
-This model is a fine-tuned version of [/disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct](https://huggingface.co//disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct) on the M^2RAR-20241225-LLM-1_10 dataset.
-
-## Model description
-
-More information needed
-
-## Intended uses & limitations
-
-More information needed
-
-## Training and evaluation data
-
-More information needed
-
-## Training procedure
-
-### Training hyperparameters
-
-The following hyperparameters were used during training:
-- learning_rate: 0.0001
-- train_batch_size: 2
-- eval_batch_size: 8
-- seed: 42
-- distributed_type: multi-GPU
-- num_devices: 2
-- gradient_accumulation_steps: 16
-- total_train_batch_size: 64
-- total_eval_batch_size: 16
-- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
-- lr_scheduler_type: cosine
-- lr_scheduler_warmup_ratio: 0.1
-- num_epochs: 3.0
-
-### Training results
-
-
-
-### Framework versions
-
-- PEFT 0.12.0
-- Transformers 4.46.1
-- Pytorch 2.3.1+cu121
-- Datasets 3.1.0
-- Tokenizers 0.20.3

llama_factory/v3-20241225-1_10/adapter_config.json

@@ -1,34 +0,0 @@
-{
-  "alpha_pattern": {},
-  "auto_mapping": null,
-  "base_model_name_or_path": "/disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct",
-  "bias": "none",
-  "fan_in_fan_out": false,
-  "inference_mode": true,
-  "init_lora_weights": true,
-  "layer_replication": null,
-  "layers_pattern": null,
-  "layers_to_transform": null,
-  "loftq_config": {},
-  "lora_alpha": 256,
-  "lora_dropout": 0.0,
-  "megatron_config": null,
-  "megatron_core": "megatron.core",
-  "modules_to_save": null,
-  "peft_type": "LORA",
-  "r": 128,
-  "rank_pattern": {},
-  "revision": null,
-  "target_modules": [
-    "up_proj",
-    "v_proj",
-    "down_proj",
-    "q_proj",
-    "o_proj",
-    "k_proj",
-    "gate_proj"
-  ],
-  "task_type": "CAUSAL_LM",
-  "use_dora": false,
-  "use_rslora": false
-}

llama_factory/v3-20241225-1_10/added_tokens.json

@@ -1,24 +0,0 @@
-{
-  "</tool_call>": 151658,
-  "<tool_call>": 151657,
-  "<|box_end|>": 151649,
-  "<|box_start|>": 151648,
-  "<|endoftext|>": 151643,
-  "<|file_sep|>": 151664,
-  "<|fim_middle|>": 151660,
-  "<|fim_pad|>": 151662,
-  "<|fim_prefix|>": 151659,
-  "<|fim_suffix|>": 151661,
-  "<|im_end|>": 151645,
-  "<|im_start|>": 151644,
-  "<|image_pad|>": 151655,
-  "<|object_ref_end|>": 151647,
-  "<|object_ref_start|>": 151646,
-  "<|quad_end|>": 151651,
-  "<|quad_start|>": 151650,
-  "<|repo_name|>": 151663,
-  "<|video_pad|>": 151656,
-  "<|vision_end|>": 151653,
-  "<|vision_pad|>": 151654,
-  "<|vision_start|>": 151652
-}

llama_factory/v3-20241225-1_10/all_results.json

@@ -1,8 +0,0 @@
-{
-    "epoch": 2.9547511312217196,
-    "total_flos": 1.6996828973608468e+18,
-    "train_loss": 1.3303501635421942,
-    "train_runtime": 7728.7219,
-    "train_samples_per_second": 0.685,
-    "train_steps_per_second": 0.01
-}

llama_factory/v3-20241225-1_10/special_tokens_map.json

@@ -1,31 +0,0 @@
-{
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>",
-    "<|object_ref_start|>",
-    "<|object_ref_end|>",
-    "<|box_start|>",
-    "<|box_end|>",
-    "<|quad_start|>",
-    "<|quad_end|>",
-    "<|vision_start|>",
-    "<|vision_end|>",
-    "<|vision_pad|>",
-    "<|image_pad|>",
-    "<|video_pad|>"
-  ],
-  "eos_token": {
-    "content": "<|im_end|>",
-    "lstrip": false,
-    "normalized": false,
-    "rstrip": false,
-    "single_word": false
-  },
-  "pad_token": {
-    "content": "<|endoftext|>",
-    "lstrip": false,
-    "normalized": false,
-    "rstrip": false,
-    "single_word": false
-  }
-}

llama_factory/v3-20241225-1_10/tokenizer_config.json

@@ -1,208 +0,0 @@
-{
-  "add_bos_token": false,
-  "add_prefix_space": false,
-  "added_tokens_decoder": {
-    "151643": {
-      "content": "<|endoftext|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151644": {
-      "content": "<|im_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151645": {
-      "content": "<|im_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151646": {
-      "content": "<|object_ref_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151647": {
-      "content": "<|object_ref_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151648": {
-      "content": "<|box_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151649": {
-      "content": "<|box_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151650": {
-      "content": "<|quad_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151651": {
-      "content": "<|quad_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151652": {
-      "content": "<|vision_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151653": {
-      "content": "<|vision_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151654": {
-      "content": "<|vision_pad|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151655": {
-      "content": "<|image_pad|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151656": {
-      "content": "<|video_pad|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151657": {
-      "content": "<tool_call>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151658": {
-      "content": "</tool_call>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151659": {
-      "content": "<|fim_prefix|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151660": {
-      "content": "<|fim_middle|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151661": {
-      "content": "<|fim_suffix|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151662": {
-      "content": "<|fim_pad|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151663": {
-      "content": "<|repo_name|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151664": {
-      "content": "<|file_sep|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    }
-  },
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>",
-    "<|object_ref_start|>",
-    "<|object_ref_end|>",
-    "<|box_start|>",
-    "<|box_end|>",
-    "<|quad_start|>",
-    "<|quad_end|>",
-    "<|vision_start|>",
-    "<|vision_end|>",
-    "<|vision_pad|>",
-    "<|image_pad|>",
-    "<|video_pad|>"
-  ],
-  "bos_token": null,
-  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
-  "clean_up_tokenization_spaces": false,
-  "eos_token": "<|im_end|>",
-  "errors": "replace",
-  "model_max_length": 131072,
-  "pad_token": "<|endoftext|>",
-  "padding_side": "right",
-  "split_special_tokens": false,
-  "tokenizer_class": "Qwen2Tokenizer",
-  "unk_token": null
-}

llama_factory/v3-20241225-1_10/train_results.json

@@ -1,8 +0,0 @@
-{
-    "epoch": 2.9547511312217196,
-    "total_flos": 1.6996828973608468e+18,
-    "train_loss": 1.3303501635421942,
-    "train_runtime": 7728.7219,
-    "train_samples_per_second": 0.685,
-    "train_steps_per_second": 0.01
-}

llama_factory/v3-20241225-1_10/trainer_log.jsonl

@@ -1,82 +0,0 @@
-{"current_steps": 1, "total_steps": 81, "loss": 1.5875, "lr": 1.1111111111111112e-05, "epoch": 0.03619909502262444, "percentage": 1.23, "elapsed_time": "0:01:42", "remaining_time": "2:16:37"}
-{"current_steps": 2, "total_steps": 81, "loss": 1.596, "lr": 2.2222222222222223e-05, "epoch": 0.07239819004524888, "percentage": 2.47, "elapsed_time": "0:03:22", "remaining_time": "2:13:20"}
-{"current_steps": 3, "total_steps": 81, "loss": 1.5499, "lr": 3.3333333333333335e-05, "epoch": 0.1085972850678733, "percentage": 3.7, "elapsed_time": "0:04:42", "remaining_time": "2:02:34"}
-{"current_steps": 4, "total_steps": 81, "loss": 1.5098, "lr": 4.4444444444444447e-05, "epoch": 0.14479638009049775, "percentage": 4.94, "elapsed_time": "0:06:22", "remaining_time": "2:02:39"}
-{"current_steps": 5, "total_steps": 81, "loss": 1.509, "lr": 5.555555555555556e-05, "epoch": 0.18099547511312217, "percentage": 6.17, "elapsed_time": "0:08:08", "remaining_time": "2:03:43"}
-{"current_steps": 6, "total_steps": 81, "loss": 1.4715, "lr": 6.666666666666667e-05, "epoch": 0.2171945701357466, "percentage": 7.41, "elapsed_time": "0:09:53", "remaining_time": "2:03:42"}
-{"current_steps": 7, "total_steps": 81, "loss": 1.4596, "lr": 7.777777777777778e-05, "epoch": 0.25339366515837103, "percentage": 8.64, "elapsed_time": "0:11:19", "remaining_time": "1:59:39"}
-{"current_steps": 8, "total_steps": 81, "loss": 1.4378, "lr": 8.888888888888889e-05, "epoch": 0.2895927601809955, "percentage": 9.88, "elapsed_time": "0:12:56", "remaining_time": "1:58:08"}
-{"current_steps": 9, "total_steps": 81, "loss": 1.4647, "lr": 0.0001, "epoch": 0.3257918552036199, "percentage": 11.11, "elapsed_time": "0:14:54", "remaining_time": "1:59:12"}
-{"current_steps": 10, "total_steps": 81, "loss": 1.4099, "lr": 9.99524110790929e-05, "epoch": 0.36199095022624433, "percentage": 12.35, "elapsed_time": "0:16:19", "remaining_time": "1:55:56"}
-{"current_steps": 11, "total_steps": 81, "loss": 1.3963, "lr": 9.980973490458728e-05, "epoch": 0.39819004524886875, "percentage": 13.58, "elapsed_time": "0:18:04", "remaining_time": "1:55:03"}
-{"current_steps": 12, "total_steps": 81, "loss": 1.4473, "lr": 9.957224306869053e-05, "epoch": 0.4343891402714932, "percentage": 14.81, "elapsed_time": "0:19:12", "remaining_time": "1:50:25"}
-{"current_steps": 13, "total_steps": 81, "loss": 1.3592, "lr": 9.924038765061042e-05, "epoch": 0.47058823529411764, "percentage": 16.05, "elapsed_time": "0:20:52", "remaining_time": "1:49:13"}
-{"current_steps": 14, "total_steps": 81, "loss": 1.3714, "lr": 9.881480035599667e-05, "epoch": 0.5067873303167421, "percentage": 17.28, "elapsed_time": "0:22:35", "remaining_time": "1:48:04"}
-{"current_steps": 15, "total_steps": 81, "loss": 1.3921, "lr": 9.829629131445342e-05, "epoch": 0.5429864253393665, "percentage": 18.52, "elapsed_time": "0:24:04", "remaining_time": "1:45:57"}
-{"current_steps": 16, "total_steps": 81, "loss": 1.364, "lr": 9.768584753741134e-05, "epoch": 0.579185520361991, "percentage": 19.75, "elapsed_time": "0:25:21", "remaining_time": "1:43:02"}
-{"current_steps": 17, "total_steps": 81, "loss": 1.3907, "lr": 9.698463103929542e-05, "epoch": 0.6153846153846154, "percentage": 20.99, "elapsed_time": "0:27:07", "remaining_time": "1:42:06"}
-{"current_steps": 18, "total_steps": 81, "loss": 1.3672, "lr": 9.619397662556435e-05, "epoch": 0.6515837104072398, "percentage": 22.22, "elapsed_time": "0:28:51", "remaining_time": "1:41:01"}
-{"current_steps": 19, "total_steps": 81, "loss": 1.3952, "lr": 9.53153893518325e-05, "epoch": 0.6877828054298643, "percentage": 23.46, "elapsed_time": "0:30:18", "remaining_time": "1:38:54"}
-{"current_steps": 20, "total_steps": 81, "loss": 1.3844, "lr": 9.435054165891109e-05, "epoch": 0.7239819004524887, "percentage": 24.69, "elapsed_time": "0:31:32", "remaining_time": "1:36:11"}
-{"current_steps": 21, "total_steps": 81, "loss": 1.3794, "lr": 9.330127018922194e-05, "epoch": 0.7601809954751131, "percentage": 25.93, "elapsed_time": "0:33:11", "remaining_time": "1:34:50"}
-{"current_steps": 22, "total_steps": 81, "loss": 1.3848, "lr": 9.21695722906443e-05, "epoch": 0.7963800904977375, "percentage": 27.16, "elapsed_time": "0:34:30", "remaining_time": "1:32:33"}
-{"current_steps": 23, "total_steps": 81, "loss": 1.3343, "lr": 9.09576022144496e-05, "epoch": 0.832579185520362, "percentage": 28.4, "elapsed_time": "0:36:14", "remaining_time": "1:31:24"}
-{"current_steps": 24, "total_steps": 81, "loss": 1.3583, "lr": 8.966766701456177e-05, "epoch": 0.8687782805429864, "percentage": 29.63, "elapsed_time": "0:37:45", "remaining_time": "1:29:40"}
-{"current_steps": 25, "total_steps": 81, "loss": 1.3213, "lr": 8.83022221559489e-05, "epoch": 0.9049773755656109, "percentage": 30.86, "elapsed_time": "0:39:36", "remaining_time": "1:28:42"}
-{"current_steps": 26, "total_steps": 81, "loss": 1.3217, "lr": 8.68638668405062e-05, "epoch": 0.9411764705882353, "percentage": 32.1, "elapsed_time": "0:41:22", "remaining_time": "1:27:31"}
-{"current_steps": 27, "total_steps": 81, "loss": 1.3899, "lr": 8.535533905932738e-05, "epoch": 0.9773755656108597, "percentage": 33.33, "elapsed_time": "0:43:01", "remaining_time": "1:26:03"}
-{"current_steps": 28, "total_steps": 81, "loss": 2.2657, "lr": 8.377951038078302e-05, "epoch": 1.0248868778280542, "percentage": 34.57, "elapsed_time": "0:44:23", "remaining_time": "1:24:01"}
-{"current_steps": 29, "total_steps": 81, "loss": 1.2028, "lr": 8.213938048432697e-05, "epoch": 1.0610859728506787, "percentage": 35.8, "elapsed_time": "0:45:55", "remaining_time": "1:22:20"}
-{"current_steps": 30, "total_steps": 81, "loss": 1.5119, "lr": 8.043807145043604e-05, "epoch": 1.0972850678733033, "percentage": 37.04, "elapsed_time": "0:48:01", "remaining_time": "1:21:38"}
-{"current_steps": 31, "total_steps": 81, "loss": 1.2249, "lr": 7.86788218175523e-05, "epoch": 1.1334841628959276, "percentage": 38.27, "elapsed_time": "0:49:18", "remaining_time": "1:19:31"}
-{"current_steps": 32, "total_steps": 81, "loss": 1.363, "lr": 7.68649804173412e-05, "epoch": 1.169683257918552, "percentage": 39.51, "elapsed_time": "0:50:47", "remaining_time": "1:17:46"}
-{"current_steps": 33, "total_steps": 81, "loss": 1.2555, "lr": 7.500000000000001e-05, "epoch": 1.2058823529411764, "percentage": 40.74, "elapsed_time": "0:52:04", "remaining_time": "1:15:45"}
-{"current_steps": 34, "total_steps": 81, "loss": 1.3404, "lr": 7.308743066175172e-05, "epoch": 1.242081447963801, "percentage": 41.98, "elapsed_time": "0:53:48", "remaining_time": "1:14:23"}
-{"current_steps": 35, "total_steps": 81, "loss": 1.1727, "lr": 7.113091308703498e-05, "epoch": 1.2782805429864252, "percentage": 43.21, "elapsed_time": "0:55:13", "remaining_time": "1:12:34"}
-{"current_steps": 36, "total_steps": 81, "loss": 1.4164, "lr": 6.91341716182545e-05, "epoch": 1.3144796380090498, "percentage": 44.44, "elapsed_time": "0:57:19", "remaining_time": "1:11:39"}
-{"current_steps": 37, "total_steps": 81, "loss": 1.2009, "lr": 6.710100716628344e-05, "epoch": 1.3506787330316743, "percentage": 45.68, "elapsed_time": "0:58:57", "remaining_time": "1:10:06"}
-{"current_steps": 38, "total_steps": 81, "loss": 1.2526, "lr": 6.503528997521366e-05, "epoch": 1.3868778280542986, "percentage": 46.91, "elapsed_time": "1:00:24", "remaining_time": "1:08:21"}
-{"current_steps": 39, "total_steps": 81, "loss": 1.2736, "lr": 6.294095225512603e-05, "epoch": 1.4230769230769231, "percentage": 48.15, "elapsed_time": "1:01:57", "remaining_time": "1:06:43"}
-{"current_steps": 40, "total_steps": 81, "loss": 1.2816, "lr": 6.0821980696905146e-05, "epoch": 1.4592760180995474, "percentage": 49.38, "elapsed_time": "1:03:23", "remaining_time": "1:04:58"}
-{"current_steps": 41, "total_steps": 81, "loss": 1.3479, "lr": 5.868240888334653e-05, "epoch": 1.495475113122172, "percentage": 50.62, "elapsed_time": "1:04:42", "remaining_time": "1:03:07"}
-{"current_steps": 42, "total_steps": 81, "loss": 1.194, "lr": 5.6526309611002594e-05, "epoch": 1.5316742081447963, "percentage": 51.85, "elapsed_time": "1:06:03", "remaining_time": "1:01:20"}
-{"current_steps": 43, "total_steps": 81, "loss": 1.3032, "lr": 5.435778713738292e-05, "epoch": 1.5678733031674208, "percentage": 53.09, "elapsed_time": "1:07:45", "remaining_time": "0:59:52"}
-{"current_steps": 44, "total_steps": 81, "loss": 1.347, "lr": 5.218096936826681e-05, "epoch": 1.6040723981900453, "percentage": 54.32, "elapsed_time": "1:09:25", "remaining_time": "0:58:22"}
-{"current_steps": 45, "total_steps": 81, "loss": 1.2273, "lr": 5e-05, "epoch": 1.6402714932126696, "percentage": 55.56, "elapsed_time": "1:10:57", "remaining_time": "0:56:46"}
-{"current_steps": 46, "total_steps": 81, "loss": 1.2478, "lr": 4.781903063173321e-05, "epoch": 1.6764705882352942, "percentage": 56.79, "elapsed_time": "1:12:20", "remaining_time": "0:55:02"}
-{"current_steps": 47, "total_steps": 81, "loss": 1.2983, "lr": 4.564221286261709e-05, "epoch": 1.7126696832579187, "percentage": 58.02, "elapsed_time": "1:13:53", "remaining_time": "0:53:27"}
-{"current_steps": 48, "total_steps": 81, "loss": 1.2293, "lr": 4.347369038899744e-05, "epoch": 1.748868778280543, "percentage": 59.26, "elapsed_time": "1:15:19", "remaining_time": "0:51:46"}
-{"current_steps": 49, "total_steps": 81, "loss": 1.2435, "lr": 4.131759111665349e-05, "epoch": 1.7850678733031673, "percentage": 60.49, "elapsed_time": "1:17:01", "remaining_time": "0:50:18"}
-{"current_steps": 50, "total_steps": 81, "loss": 1.3395, "lr": 3.917801930309486e-05, "epoch": 1.8212669683257918, "percentage": 61.73, "elapsed_time": "1:18:55", "remaining_time": "0:48:56"}
-{"current_steps": 51, "total_steps": 81, "loss": 1.2179, "lr": 3.705904774487396e-05, "epoch": 1.8574660633484164, "percentage": 62.96, "elapsed_time": "1:20:47", "remaining_time": "0:47:31"}
-{"current_steps": 52, "total_steps": 81, "loss": 1.3069, "lr": 3.4964710024786354e-05, "epoch": 1.8936651583710407, "percentage": 64.2, "elapsed_time": "1:22:37", "remaining_time": "0:46:04"}
-{"current_steps": 53, "total_steps": 81, "loss": 1.2125, "lr": 3.289899283371657e-05, "epoch": 1.9298642533936652, "percentage": 65.43, "elapsed_time": "1:24:04", "remaining_time": "0:44:24"}
-{"current_steps": 54, "total_steps": 81, "loss": 1.2311, "lr": 3.086582838174551e-05, "epoch": 1.9660633484162897, "percentage": 66.67, "elapsed_time": "1:25:38", "remaining_time": "0:42:49"}
-{"current_steps": 55, "total_steps": 81, "loss": 2.1265, "lr": 2.886908691296504e-05, "epoch": 2.013574660633484, "percentage": 67.9, "elapsed_time": "1:27:28", "remaining_time": "0:41:21"}
-{"current_steps": 56, "total_steps": 81, "loss": 1.2162, "lr": 2.6912569338248315e-05, "epoch": 2.0497737556561084, "percentage": 69.14, "elapsed_time": "1:29:01", "remaining_time": "0:39:44"}
-{"current_steps": 57, "total_steps": 81, "loss": 1.2231, "lr": 2.500000000000001e-05, "epoch": 2.085972850678733, "percentage": 70.37, "elapsed_time": "1:30:44", "remaining_time": "0:38:12"}
-{"current_steps": 58, "total_steps": 81, "loss": 1.2214, "lr": 2.3135019582658802e-05, "epoch": 2.1221719457013575, "percentage": 71.6, "elapsed_time": "1:32:26", "remaining_time": "0:36:39"}
-{"current_steps": 59, "total_steps": 81, "loss": 1.2256, "lr": 2.132117818244771e-05, "epoch": 2.158371040723982, "percentage": 72.84, "elapsed_time": "1:34:08", "remaining_time": "0:35:06"}
-{"current_steps": 60, "total_steps": 81, "loss": 1.0538, "lr": 1.9561928549563968e-05, "epoch": 2.1945701357466065, "percentage": 74.07, "elapsed_time": "1:35:22", "remaining_time": "0:33:22"}
-{"current_steps": 61, "total_steps": 81, "loss": 1.2867, "lr": 1.7860619515673033e-05, "epoch": 2.230769230769231, "percentage": 75.31, "elapsed_time": "1:37:14", "remaining_time": "0:31:52"}
-{"current_steps": 62, "total_steps": 81, "loss": 1.2971, "lr": 1.622048961921699e-05, "epoch": 2.266968325791855, "percentage": 76.54, "elapsed_time": "1:38:48", "remaining_time": "0:30:16"}
-{"current_steps": 63, "total_steps": 81, "loss": 1.1491, "lr": 1.4644660940672627e-05, "epoch": 2.3031674208144794, "percentage": 77.78, "elapsed_time": "1:40:25", "remaining_time": "0:28:41"}
-{"current_steps": 64, "total_steps": 81, "loss": 1.2826, "lr": 1.3136133159493802e-05, "epoch": 2.339366515837104, "percentage": 79.01, "elapsed_time": "1:42:30", "remaining_time": "0:27:13"}
-{"current_steps": 65, "total_steps": 81, "loss": 1.2968, "lr": 1.1697777844051105e-05, "epoch": 2.3755656108597285, "percentage": 80.25, "elapsed_time": "1:44:19", "remaining_time": "0:25:40"}
-{"current_steps": 66, "total_steps": 81, "loss": 1.129, "lr": 1.0332332985438248e-05, "epoch": 2.411764705882353, "percentage": 81.48, "elapsed_time": "1:45:28", "remaining_time": "0:23:58"}
-{"current_steps": 67, "total_steps": 81, "loss": 1.3448, "lr": 9.042397785550405e-06, "epoch": 2.4479638009049776, "percentage": 82.72, "elapsed_time": "1:47:12", "remaining_time": "0:22:24"}
-{"current_steps": 68, "total_steps": 81, "loss": 1.2397, "lr": 7.830427709355725e-06, "epoch": 2.484162895927602, "percentage": 83.95, "elapsed_time": "1:48:38", "remaining_time": "0:20:46"}
-{"current_steps": 69, "total_steps": 81, "loss": 1.1677, "lr": 6.698729810778065e-06, "epoch": 2.520361990950226, "percentage": 85.19, "elapsed_time": "1:50:10", "remaining_time": "0:19:09"}
-{"current_steps": 70, "total_steps": 81, "loss": 1.1872, "lr": 5.649458341088915e-06, "epoch": 2.5565610859728505, "percentage": 86.42, "elapsed_time": "1:51:20", "remaining_time": "0:17:29"}
-{"current_steps": 71, "total_steps": 81, "loss": 1.1483, "lr": 4.684610648167503e-06, "epoch": 2.5927601809954752, "percentage": 87.65, "elapsed_time": "1:53:03", "remaining_time": "0:15:55"}
-{"current_steps": 72, "total_steps": 81, "loss": 1.2921, "lr": 3.8060233744356633e-06, "epoch": 2.6289592760180995, "percentage": 88.89, "elapsed_time": "1:54:23", "remaining_time": "0:14:17"}
-{"current_steps": 73, "total_steps": 81, "loss": 1.2214, "lr": 3.0153689607045845e-06, "epoch": 2.665158371040724, "percentage": 90.12, "elapsed_time": "1:56:11", "remaining_time": "0:12:43"}
-{"current_steps": 74, "total_steps": 81, "loss": 1.1808, "lr": 2.314152462588659e-06, "epoch": 2.7013574660633486, "percentage": 91.36, "elapsed_time": "1:57:54", "remaining_time": "0:11:09"}
-{"current_steps": 75, "total_steps": 81, "loss": 1.2334, "lr": 1.70370868554659e-06, "epoch": 2.737556561085973, "percentage": 92.59, "elapsed_time": "1:59:37", "remaining_time": "0:09:34"}
-{"current_steps": 76, "total_steps": 81, "loss": 1.2575, "lr": 1.1851996440033319e-06, "epoch": 2.773755656108597, "percentage": 93.83, "elapsed_time": "2:01:14", "remaining_time": "0:07:58"}
-{"current_steps": 77, "total_steps": 81, "loss": 1.3336, "lr": 7.596123493895991e-07, "epoch": 2.8099547511312215, "percentage": 95.06, "elapsed_time": "2:02:34", "remaining_time": "0:06:22"}
-{"current_steps": 78, "total_steps": 81, "loss": 1.0837, "lr": 4.277569313094809e-07, "epoch": 2.8461538461538463, "percentage": 96.3, "elapsed_time": "2:04:06", "remaining_time": "0:04:46"}
-{"current_steps": 79, "total_steps": 81, "loss": 1.2523, "lr": 1.9026509541272275e-07, "epoch": 2.8823529411764706, "percentage": 97.53, "elapsed_time": "2:05:38", "remaining_time": "0:03:10"}
-{"current_steps": 80, "total_steps": 81, "loss": 1.2335, "lr": 4.7588920907110094e-08, "epoch": 2.918552036199095, "percentage": 98.77, "elapsed_time": "2:07:06", "remaining_time": "0:01:35"}
-{"current_steps": 81, "total_steps": 81, "loss": 1.2131, "lr": 0.0, "epoch": 2.9547511312217196, "percentage": 100.0, "elapsed_time": "2:08:13", "remaining_time": "0:00:00"}
-{"current_steps": 81, "total_steps": 81, "epoch": 2.9547511312217196, "percentage": 100.0, "elapsed_time": "2:08:48", "remaining_time": "0:00:00"}

llama_factory/v3-20241225-1_10/trainer_state.json

@@ -1,609 +0,0 @@
-{
-  "best_metric": null,
-  "best_model_checkpoint": null,
-  "epoch": 2.9547511312217196,
-  "eval_steps": 500,
-  "global_step": 81,
-  "is_hyper_param_search": false,
-  "is_local_process_zero": true,
-  "is_world_process_zero": true,
-  "log_history": [
-    {
-      "epoch": 0.03619909502262444,
-      "grad_norm": 1.3874496221542358,
-      "learning_rate": 1.1111111111111112e-05,
-      "loss": 1.5875,
-      "step": 1
-    },
-    {
-      "epoch": 0.07239819004524888,
-      "grad_norm": 1.405678391456604,
-      "learning_rate": 2.2222222222222223e-05,
-      "loss": 1.596,
-      "step": 2
-    },
-    {
-      "epoch": 0.1085972850678733,
-      "grad_norm": 1.303895354270935,
-      "learning_rate": 3.3333333333333335e-05,
-      "loss": 1.5499,
-      "step": 3
-    },
-    {
-      "epoch": 0.14479638009049775,
-      "grad_norm": 0.6614035367965698,
-      "learning_rate": 4.4444444444444447e-05,
-      "loss": 1.5098,
-      "step": 4
-    },
-    {
-      "epoch": 0.18099547511312217,
-      "grad_norm": 0.36806759238243103,
-      "learning_rate": 5.555555555555556e-05,
-      "loss": 1.509,
-      "step": 5
-    },
-    {
-      "epoch": 0.2171945701357466,
-      "grad_norm": 0.49412667751312256,
-      "learning_rate": 6.666666666666667e-05,
-      "loss": 1.4715,
-      "step": 6
-    },
-    {
-      "epoch": 0.25339366515837103,
-      "grad_norm": 0.536666989326477,
-      "learning_rate": 7.777777777777778e-05,
-      "loss": 1.4596,
-      "step": 7
-    },
-    {
-      "epoch": 0.2895927601809955,
-      "grad_norm": 0.4359913468360901,
-      "learning_rate": 8.888888888888889e-05,
-      "loss": 1.4378,
-      "step": 8
-    },
-    {
-      "epoch": 0.3257918552036199,
-      "grad_norm": 0.33356979489326477,
-      "learning_rate": 0.0001,
-      "loss": 1.4647,
-      "step": 9
-    },
-    {
-      "epoch": 0.36199095022624433,
-      "grad_norm": 0.22868072986602783,
-      "learning_rate": 9.99524110790929e-05,
-      "loss": 1.4099,
-      "step": 10
-    },
-    {
-      "epoch": 0.39819004524886875,
-      "grad_norm": 0.24769297242164612,
-      "learning_rate": 9.980973490458728e-05,
-      "loss": 1.3963,
-      "step": 11
-    },
-    {
-      "epoch": 0.4343891402714932,
-      "grad_norm": 0.3057805597782135,
-      "learning_rate": 9.957224306869053e-05,
-      "loss": 1.4473,
-      "step": 12
-    },
-    {
-      "epoch": 0.47058823529411764,
-      "grad_norm": 0.21798226237297058,
-      "learning_rate": 9.924038765061042e-05,
-      "loss": 1.3592,
-      "step": 13
-    },
-    {
-      "epoch": 0.5067873303167421,
-      "grad_norm": 0.180103600025177,
-      "learning_rate": 9.881480035599667e-05,
-      "loss": 1.3714,
-      "step": 14
-    },
-    {
-      "epoch": 0.5429864253393665,
-      "grad_norm": 0.20826148986816406,
-      "learning_rate": 9.829629131445342e-05,
-      "loss": 1.3921,
-      "step": 15
-    },
-    {
-      "epoch": 0.579185520361991,
-      "grad_norm": 0.22720584273338318,
-      "learning_rate": 9.768584753741134e-05,
-      "loss": 1.364,
-      "step": 16
-    },
-    {
-      "epoch": 0.6153846153846154,
-      "grad_norm": 0.18786540627479553,
-      "learning_rate": 9.698463103929542e-05,
-      "loss": 1.3907,
-      "step": 17
-    },
-    {
-      "epoch": 0.6515837104072398,
-      "grad_norm": 0.17362689971923828,
-      "learning_rate": 9.619397662556435e-05,
-      "loss": 1.3672,
-      "step": 18
-    },
-    {
-      "epoch": 0.6877828054298643,
-      "grad_norm": 0.1910216510295868,
-      "learning_rate": 9.53153893518325e-05,
-      "loss": 1.3952,
-      "step": 19
-    },
-    {
-      "epoch": 0.7239819004524887,
-      "grad_norm": 0.1894685924053192,
-      "learning_rate": 9.435054165891109e-05,
-      "loss": 1.3844,
-      "step": 20
-    },
-    {
-      "epoch": 0.7601809954751131,
-      "grad_norm": 0.17293618619441986,
-      "learning_rate": 9.330127018922194e-05,
-      "loss": 1.3794,
-      "step": 21
-    },
-    {
-      "epoch": 0.7963800904977375,
-      "grad_norm": 0.17102670669555664,
-      "learning_rate": 9.21695722906443e-05,
-      "loss": 1.3848,
-      "step": 22
-    },
-    {
-      "epoch": 0.832579185520362,
-      "grad_norm": 0.15698343515396118,
-      "learning_rate": 9.09576022144496e-05,
-      "loss": 1.3343,
-      "step": 23
-    },
-    {
-      "epoch": 0.8687782805429864,
-      "grad_norm": 0.17238566279411316,
-      "learning_rate": 8.966766701456177e-05,
-      "loss": 1.3583,
-      "step": 24
-    },
-    {
-      "epoch": 0.9049773755656109,
-      "grad_norm": 0.1613931804895401,
-      "learning_rate": 8.83022221559489e-05,
-      "loss": 1.3213,
-      "step": 25
-    },
-    {
-      "epoch": 0.9411764705882353,
-      "grad_norm": 0.15348491072654724,
-      "learning_rate": 8.68638668405062e-05,
-      "loss": 1.3217,
-      "step": 26
-    },
-    {
-      "epoch": 0.9773755656108597,
-      "grad_norm": 0.1582302302122116,
-      "learning_rate": 8.535533905932738e-05,
-      "loss": 1.3899,
-      "step": 27
-    },
-    {
-      "epoch": 1.0248868778280542,
-      "grad_norm": 0.30110427737236023,
-      "learning_rate": 8.377951038078302e-05,
-      "loss": 2.2657,
-      "step": 28
-    },
-    {
-      "epoch": 1.0610859728506787,
-      "grad_norm": 0.1470889151096344,
-      "learning_rate": 8.213938048432697e-05,
-      "loss": 1.2028,
-      "step": 29
-    },
-    {
-      "epoch": 1.0972850678733033,
-      "grad_norm": 0.16698622703552246,
-      "learning_rate": 8.043807145043604e-05,
-      "loss": 1.5119,
-      "step": 30
-    },
-    {
-      "epoch": 1.1334841628959276,
-      "grad_norm": 0.15491284430027008,
-      "learning_rate": 7.86788218175523e-05,
-      "loss": 1.2249,
-      "step": 31
-    },
-    {
-      "epoch": 1.169683257918552,
-      "grad_norm": 0.16662749648094177,
-      "learning_rate": 7.68649804173412e-05,
-      "loss": 1.363,
-      "step": 32
-    },
-    {
-      "epoch": 1.2058823529411764,
-      "grad_norm": 0.15660426020622253,
-      "learning_rate": 7.500000000000001e-05,
-      "loss": 1.2555,
-      "step": 33
-    },
-    {
-      "epoch": 1.242081447963801,
-      "grad_norm": 0.15860170125961304,
-      "learning_rate": 7.308743066175172e-05,
-      "loss": 1.3404,
-      "step": 34
-    },
-    {
-      "epoch": 1.2782805429864252,
-      "grad_norm": 0.1499716341495514,
-      "learning_rate": 7.113091308703498e-05,
-      "loss": 1.1727,
-      "step": 35
-    },
-    {
-      "epoch": 1.3144796380090498,
-      "grad_norm": 0.17482785880565643,
-      "learning_rate": 6.91341716182545e-05,
-      "loss": 1.4164,
-      "step": 36
-    },
-    {
-      "epoch": 1.3506787330316743,
-      "grad_norm": 0.15533021092414856,
-      "learning_rate": 6.710100716628344e-05,
-      "loss": 1.2009,
-      "step": 37
-    },
-    {
-      "epoch": 1.3868778280542986,
-      "grad_norm": 0.15700975060462952,
-      "learning_rate": 6.503528997521366e-05,
-      "loss": 1.2526,
-      "step": 38
-    },
-    {
-      "epoch": 1.4230769230769231,
-      "grad_norm": 0.1649225950241089,
-      "learning_rate": 6.294095225512603e-05,
-      "loss": 1.2736,
-      "step": 39
-    },
-    {
-      "epoch": 1.4592760180995474,
-      "grad_norm": 0.1609291434288025,
-      "learning_rate": 6.0821980696905146e-05,
-      "loss": 1.2816,
-      "step": 40
-    },
-    {
-      "epoch": 1.495475113122172,
-      "grad_norm": 0.1757732331752777,
-      "learning_rate": 5.868240888334653e-05,
-      "loss": 1.3479,
-      "step": 41
-    },
-    {
-      "epoch": 1.5316742081447963,
-      "grad_norm": 0.16721291840076447,
-      "learning_rate": 5.6526309611002594e-05,
-      "loss": 1.194,
-      "step": 42
-    },
-    {
-      "epoch": 1.5678733031674208,
-      "grad_norm": 0.1664680391550064,
-      "learning_rate": 5.435778713738292e-05,
-      "loss": 1.3032,
-      "step": 43
-    },
-    {
-      "epoch": 1.6040723981900453,
-      "grad_norm": 0.16910715401172638,
-      "learning_rate": 5.218096936826681e-05,
-      "loss": 1.347,
-      "step": 44
-    },
-    {
-      "epoch": 1.6402714932126696,
-      "grad_norm": 0.17370687425136566,
-      "learning_rate": 5e-05,
-      "loss": 1.2273,
-      "step": 45
-    },
-    {
-      "epoch": 1.6764705882352942,
-      "grad_norm": 0.16753946244716644,
-      "learning_rate": 4.781903063173321e-05,
-      "loss": 1.2478,
-      "step": 46
-    },
-    {
-      "epoch": 1.7126696832579187,
-      "grad_norm": 0.16744458675384521,
-      "learning_rate": 4.564221286261709e-05,
-      "loss": 1.2983,
-      "step": 47
-    },
-    {
-      "epoch": 1.748868778280543,
-      "grad_norm": 0.15938393771648407,
-      "learning_rate": 4.347369038899744e-05,
-      "loss": 1.2293,
-      "step": 48
-    },
-    {
-      "epoch": 1.7850678733031673,
-      "grad_norm": 0.1637125313282013,
-      "learning_rate": 4.131759111665349e-05,
-      "loss": 1.2435,
-      "step": 49
-    },
-    {
-      "epoch": 1.8212669683257918,
-      "grad_norm": 0.1678181141614914,
-      "learning_rate": 3.917801930309486e-05,
-      "loss": 1.3395,
-      "step": 50
-    },
-    {
-      "epoch": 1.8574660633484164,
-      "grad_norm": 0.16167080402374268,
-      "learning_rate": 3.705904774487396e-05,
-      "loss": 1.2179,
-      "step": 51
-    },
-    {
-      "epoch": 1.8936651583710407,
-      "grad_norm": 0.17365491390228271,
-      "learning_rate": 3.4964710024786354e-05,
-      "loss": 1.3069,
-      "step": 52
-    },
-    {
-      "epoch": 1.9298642533936652,
-      "grad_norm": 0.16826802492141724,
-      "learning_rate": 3.289899283371657e-05,
-      "loss": 1.2125,
-      "step": 53
-    },
-    {
-      "epoch": 1.9660633484162897,
-      "grad_norm": 0.16662856936454773,
-      "learning_rate": 3.086582838174551e-05,
-      "loss": 1.2311,
-      "step": 54
-    },
-    {
-      "epoch": 2.013574660633484,
-      "grad_norm": 0.33721455931663513,
-      "learning_rate": 2.886908691296504e-05,
-      "loss": 2.1265,
-      "step": 55
-    },
-    {
-      "epoch": 2.0497737556561084,
-      "grad_norm": 0.17296265065670013,
-      "learning_rate": 2.6912569338248315e-05,
-      "loss": 1.2162,
-      "step": 56
-    },
-    {
-      "epoch": 2.085972850678733,
-      "grad_norm": 0.16787242889404297,
-      "learning_rate": 2.500000000000001e-05,
-      "loss": 1.2231,
-      "step": 57
-    },
-    {
-      "epoch": 2.1221719457013575,
-      "grad_norm": 0.17678987979888916,
-      "learning_rate": 2.3135019582658802e-05,
-      "loss": 1.2214,
-      "step": 58
-    },
-    {
-      "epoch": 2.158371040723982,
-      "grad_norm": 0.17289990186691284,
-      "learning_rate": 2.132117818244771e-05,
-      "loss": 1.2256,
-      "step": 59
-    },
-    {
-      "epoch": 2.1945701357466065,
-      "grad_norm": 0.15937398374080658,
-      "learning_rate": 1.9561928549563968e-05,
-      "loss": 1.0538,
-      "step": 60
-    },
-    {
-      "epoch": 2.230769230769231,
-      "grad_norm": 0.1900373250246048,
-      "learning_rate": 1.7860619515673033e-05,
-      "loss": 1.2867,
-      "step": 61
-    },
-    {
-      "epoch": 2.266968325791855,
-      "grad_norm": 0.20580348372459412,
-      "learning_rate": 1.622048961921699e-05,
-      "loss": 1.2971,
-      "step": 62
-    },
-    {
-      "epoch": 2.3031674208144794,
-      "grad_norm": 0.17183683812618256,
-      "learning_rate": 1.4644660940672627e-05,
-      "loss": 1.1491,
-      "step": 63
-    },
-    {
-      "epoch": 2.339366515837104,
-      "grad_norm": 0.17931026220321655,
-      "learning_rate": 1.3136133159493802e-05,
-      "loss": 1.2826,
-      "step": 64
-    },
-    {
-      "epoch": 2.3755656108597285,
-      "grad_norm": 0.18058271706104279,
-      "learning_rate": 1.1697777844051105e-05,
-      "loss": 1.2968,
-      "step": 65
-    },
-    {
-      "epoch": 2.411764705882353,
-      "grad_norm": 0.1691102683544159,
-      "learning_rate": 1.0332332985438248e-05,
-      "loss": 1.129,
-      "step": 66
-    },
-    {
-      "epoch": 2.4479638009049776,
-      "grad_norm": 0.17673338949680328,
-      "learning_rate": 9.042397785550405e-06,
-      "loss": 1.3448,
-      "step": 67
-    },
-    {
-      "epoch": 2.484162895927602,
-      "grad_norm": 0.17145270109176636,
-      "learning_rate": 7.830427709355725e-06,
-      "loss": 1.2397,
-      "step": 68
-    },
-    {
-      "epoch": 2.520361990950226,
-      "grad_norm": 0.17466092109680176,
-      "learning_rate": 6.698729810778065e-06,
-      "loss": 1.1677,
-      "step": 69
-    },
-    {
-      "epoch": 2.5565610859728505,
-      "grad_norm": 0.17480231821537018,
-      "learning_rate": 5.649458341088915e-06,
-      "loss": 1.1872,
-      "step": 70
-    },
-    {
-      "epoch": 2.5927601809954752,
-      "grad_norm": 0.1793895661830902,
-      "learning_rate": 4.684610648167503e-06,
-      "loss": 1.1483,
-      "step": 71
-    },
-    {
-      "epoch": 2.6289592760180995,
-      "grad_norm": 0.1792633980512619,
-      "learning_rate": 3.8060233744356633e-06,
-      "loss": 1.2921,
-      "step": 72
-    },
-    {
-      "epoch": 2.665158371040724,
-      "grad_norm": 0.16958290338516235,
-      "learning_rate": 3.0153689607045845e-06,
-      "loss": 1.2214,
-      "step": 73
-    },
-    {
-      "epoch": 2.7013574660633486,
-      "grad_norm": 0.16733671724796295,
-      "learning_rate": 2.314152462588659e-06,
-      "loss": 1.1808,
-      "step": 74
-    },
-    {
-      "epoch": 2.737556561085973,
-      "grad_norm": 0.18359480798244476,
-      "learning_rate": 1.70370868554659e-06,
-      "loss": 1.2334,
-      "step": 75
-    },
-    {
-      "epoch": 2.773755656108597,
-      "grad_norm": 0.1765512228012085,
-      "learning_rate": 1.1851996440033319e-06,
-      "loss": 1.2575,
-      "step": 76
-    },
-    {
-      "epoch": 2.8099547511312215,
-      "grad_norm": 0.19382162392139435,
-      "learning_rate": 7.596123493895991e-07,
-      "loss": 1.3336,
-      "step": 77
-    },
-    {
-      "epoch": 2.8461538461538463,
-      "grad_norm": 0.16584135591983795,
-      "learning_rate": 4.277569313094809e-07,
-      "loss": 1.0837,
-      "step": 78
-    },
-    {
-      "epoch": 2.8823529411764706,
-      "grad_norm": 0.17709200084209442,
-      "learning_rate": 1.9026509541272275e-07,
-      "loss": 1.2523,
-      "step": 79
-    },
-    {
-      "epoch": 2.918552036199095,
-      "grad_norm": 0.17595721781253815,
-      "learning_rate": 4.7588920907110094e-08,
-      "loss": 1.2335,
-      "step": 80
-    },
-    {
-      "epoch": 2.9547511312217196,
-      "grad_norm": 0.18491940200328827,
-      "learning_rate": 0.0,
-      "loss": 1.2131,
-      "step": 81
-    },
-    {
-      "epoch": 2.9547511312217196,
-      "step": 81,
-      "total_flos": 1.6996828973608468e+18,
-      "train_loss": 1.3303501635421942,
-      "train_runtime": 7728.7219,
-      "train_samples_per_second": 0.685,
-      "train_steps_per_second": 0.01
-    }
-  ],
-  "logging_steps": 1,
-  "max_steps": 81,
-  "num_input_tokens_seen": 0,
-  "num_train_epochs": 3,
-  "save_steps": 200,
-  "stateful_callbacks": {
-    "TrainerControl": {
-      "args": {
-        "should_epoch_stop": false,
-        "should_evaluate": false,
-        "should_log": false,
-        "should_save": true,
-        "should_training_stop": true
-      },
-      "attributes": {}
-    }
-  },
-  "total_flos": 1.6996828973608468e+18,
-  "train_batch_size": 2,
-  "trial_name": null,
-  "trial_params": null
-}

llama_factory/v3-20241225-1_5/README.md

@@ -1,62 +0,0 @@
----
-library_name: peft
-license: other
-base_model: /disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct
-tags:
-- llama-factory
-- lora
-- generated_from_trainer
-model-index:
-- name: v3-20241225-1_5
-  results: []
----
-
-<!-- This model card has been generated automatically according to the information the Trainer had access to. You
-should probably proofread and complete it, then remove this comment. -->
-
-# v3-20241225-1_5
-
-This model is a fine-tuned version of [/disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct](https://huggingface.co//disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct) on the M^2RAR-20241225-LLM-1_5 dataset.
-
-## Model description
-
-More information needed
-
-## Intended uses & limitations
-
-More information needed
-
-## Training and evaluation data
-
-More information needed
-
-## Training procedure
-
-### Training hyperparameters
-
-The following hyperparameters were used during training:
-- learning_rate: 0.0001
-- train_batch_size: 2
-- eval_batch_size: 8
-- seed: 42
-- distributed_type: multi-GPU
-- num_devices: 2
-- gradient_accumulation_steps: 16
-- total_train_batch_size: 64
-- total_eval_batch_size: 16
-- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
-- lr_scheduler_type: cosine
-- lr_scheduler_warmup_ratio: 0.1
-- num_epochs: 3.0
-
-### Training results
-
-
-
-### Framework versions
-
-- PEFT 0.12.0
-- Transformers 4.46.1
-- Pytorch 2.3.1+cu121
-- Datasets 3.1.0
-- Tokenizers 0.20.3

llama_factory/v3-20241225-1_5/adapter_config.json

@@ -1,34 +0,0 @@
-{
-  "alpha_pattern": {},
-  "auto_mapping": null,
-  "base_model_name_or_path": "/disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct",
-  "bias": "none",
-  "fan_in_fan_out": false,
-  "inference_mode": true,
-  "init_lora_weights": true,
-  "layer_replication": null,
-  "layers_pattern": null,
-  "layers_to_transform": null,
-  "loftq_config": {},
-  "lora_alpha": 256,
-  "lora_dropout": 0.0,
-  "megatron_config": null,
-  "megatron_core": "megatron.core",
-  "modules_to_save": null,
-  "peft_type": "LORA",
-  "r": 128,
-  "rank_pattern": {},
-  "revision": null,
-  "target_modules": [
-    "q_proj",
-    "gate_proj",
-    "k_proj",
-    "down_proj",
-    "v_proj",
-    "o_proj",
-    "up_proj"
-  ],
-  "task_type": "CAUSAL_LM",
-  "use_dora": false,
-  "use_rslora": false
-}

llama_factory/v3-20241225-1_5/added_tokens.json

@@ -1,24 +0,0 @@
-{
-  "</tool_call>": 151658,
-  "<tool_call>": 151657,
-  "<|box_end|>": 151649,
-  "<|box_start|>": 151648,
-  "<|endoftext|>": 151643,
-  "<|file_sep|>": 151664,
-  "<|fim_middle|>": 151660,
-  "<|fim_pad|>": 151662,
-  "<|fim_prefix|>": 151659,
-  "<|fim_suffix|>": 151661,
-  "<|im_end|>": 151645,
-  "<|im_start|>": 151644,
-  "<|image_pad|>": 151655,
-  "<|object_ref_end|>": 151647,
-  "<|object_ref_start|>": 151646,
-  "<|quad_end|>": 151651,
-  "<|quad_start|>": 151650,
-  "<|repo_name|>": 151663,
-  "<|video_pad|>": 151656,
-  "<|vision_end|>": 151653,
-  "<|vision_pad|>": 151654,
-  "<|vision_start|>": 151652
-}

llama_factory/v3-20241225-1_5/all_results.json

@@ -1,8 +0,0 @@
-{
-    "epoch": 2.987551867219917,
-    "total_flos": 1.9253998183608484e+18,
-    "train_loss": 1.3520980450842115,
-    "train_runtime": 8756.8894,
-    "train_samples_per_second": 0.659,
-    "train_steps_per_second": 0.01
-}

llama_factory/v3-20241225-1_5/special_tokens_map.json

@@ -1,31 +0,0 @@
-{
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>",
-    "<|object_ref_start|>",
-    "<|object_ref_end|>",
-    "<|box_start|>",
-    "<|box_end|>",
-    "<|quad_start|>",
-    "<|quad_end|>",
-    "<|vision_start|>",
-    "<|vision_end|>",
-    "<|vision_pad|>",
-    "<|image_pad|>",
-    "<|video_pad|>"
-  ],
-  "eos_token": {
-    "content": "<|im_end|>",
-    "lstrip": false,
-    "normalized": false,
-    "rstrip": false,
-    "single_word": false
-  },
-  "pad_token": {
-    "content": "<|endoftext|>",
-    "lstrip": false,
-    "normalized": false,
-    "rstrip": false,
-    "single_word": false
-  }
-}

llama_factory/v3-20241225-1_5/tokenizer_config.json

@@ -1,208 +0,0 @@
-{
-  "add_bos_token": false,
-  "add_prefix_space": false,
-  "added_tokens_decoder": {
-    "151643": {
-      "content": "<|endoftext|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151644": {
-      "content": "<|im_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151645": {
-      "content": "<|im_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151646": {
-      "content": "<|object_ref_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151647": {
-      "content": "<|object_ref_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151648": {
-      "content": "<|box_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151649": {
-      "content": "<|box_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151650": {
-      "content": "<|quad_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151651": {
-      "content": "<|quad_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151652": {
-      "content": "<|vision_start|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151653": {
-      "content": "<|vision_end|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151654": {
-      "content": "<|vision_pad|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151655": {
-      "content": "<|image_pad|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151656": {
-      "content": "<|video_pad|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": true
-    },
-    "151657": {
-      "content": "<tool_call>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151658": {
-      "content": "</tool_call>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151659": {
-      "content": "<|fim_prefix|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151660": {
-      "content": "<|fim_middle|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151661": {
-      "content": "<|fim_suffix|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151662": {
-      "content": "<|fim_pad|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151663": {
-      "content": "<|repo_name|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    },
-    "151664": {
-      "content": "<|file_sep|>",
-      "lstrip": false,
-      "normalized": false,
-      "rstrip": false,
-      "single_word": false,
-      "special": false
-    }
-  },
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>",
-    "<|object_ref_start|>",
-    "<|object_ref_end|>",
-    "<|box_start|>",
-    "<|box_end|>",
-    "<|quad_start|>",
-    "<|quad_end|>",
-    "<|vision_start|>",
-    "<|vision_end|>",
-    "<|vision_pad|>",
-    "<|image_pad|>",
-    "<|video_pad|>"
-  ],
-  "bos_token": null,
-  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
-  "clean_up_tokenization_spaces": false,
-  "eos_token": "<|im_end|>",
-  "errors": "replace",
-  "model_max_length": 131072,
-  "pad_token": "<|endoftext|>",
-  "padding_side": "right",
-  "split_special_tokens": false,
-  "tokenizer_class": "Qwen2Tokenizer",
-  "unk_token": null
-}

llama_factory/v3-20241225-1_5/train_results.json

@@ -1,8 +0,0 @@
-{
-    "epoch": 2.987551867219917,
-    "total_flos": 1.9253998183608484e+18,
-    "train_loss": 1.3520980450842115,
-    "train_runtime": 8756.8894,
-    "train_samples_per_second": 0.659,
-    "train_steps_per_second": 0.01
-}

llama_factory/v3-20241225-1_5/trainer_log.jsonl

@@ -1,91 +0,0 @@
-{"current_steps": 1, "total_steps": 90, "loss": 1.6367, "lr": 1.1111111111111112e-05, "epoch": 0.03319502074688797, "percentage": 1.11, "elapsed_time": "0:01:29", "remaining_time": "2:12:33"}
-{"current_steps": 2, "total_steps": 90, "loss": 1.5947, "lr": 2.2222222222222223e-05, "epoch": 0.06639004149377593, "percentage": 2.22, "elapsed_time": "0:03:25", "remaining_time": "2:30:54"}
-{"current_steps": 3, "total_steps": 90, "loss": 1.6033, "lr": 3.3333333333333335e-05, "epoch": 0.0995850622406639, "percentage": 3.33, "elapsed_time": "0:05:22", "remaining_time": "2:35:45"}
-{"current_steps": 4, "total_steps": 90, "loss": 1.4665, "lr": 4.4444444444444447e-05, "epoch": 0.13278008298755187, "percentage": 4.44, "elapsed_time": "0:07:24", "remaining_time": "2:39:20"}
-{"current_steps": 5, "total_steps": 90, "loss": 1.4843, "lr": 5.555555555555556e-05, "epoch": 0.16597510373443983, "percentage": 5.56, "elapsed_time": "0:08:54", "remaining_time": "2:31:20"}
-{"current_steps": 6, "total_steps": 90, "loss": 1.5095, "lr": 6.666666666666667e-05, "epoch": 0.1991701244813278, "percentage": 6.67, "elapsed_time": "0:10:22", "remaining_time": "2:25:12"}
-{"current_steps": 7, "total_steps": 90, "loss": 1.5704, "lr": 7.777777777777778e-05, "epoch": 0.23236514522821577, "percentage": 7.78, "elapsed_time": "0:12:05", "remaining_time": "2:23:25"}
-{"current_steps": 8, "total_steps": 90, "loss": 1.4808, "lr": 8.888888888888889e-05, "epoch": 0.26556016597510373, "percentage": 8.89, "elapsed_time": "0:13:39", "remaining_time": "2:19:59"}
-{"current_steps": 9, "total_steps": 90, "loss": 1.4379, "lr": 0.0001, "epoch": 0.2987551867219917, "percentage": 10.0, "elapsed_time": "0:14:58", "remaining_time": "2:14:42"}
-{"current_steps": 10, "total_steps": 90, "loss": 1.3778, "lr": 9.996239762521151e-05, "epoch": 0.33195020746887965, "percentage": 11.11, "elapsed_time": "0:16:50", "remaining_time": "2:14:47"}
-{"current_steps": 11, "total_steps": 90, "loss": 1.4325, "lr": 9.98496470583896e-05, "epoch": 0.3651452282157676, "percentage": 12.22, "elapsed_time": "0:18:29", "remaining_time": "2:12:51"}
-{"current_steps": 12, "total_steps": 90, "loss": 1.4703, "lr": 9.966191788709716e-05, "epoch": 0.3983402489626556, "percentage": 13.33, "elapsed_time": "0:20:11", "remaining_time": "2:11:11"}
-{"current_steps": 13, "total_steps": 90, "loss": 1.4177, "lr": 9.939949247384046e-05, "epoch": 0.4315352697095436, "percentage": 14.44, "elapsed_time": "0:21:50", "remaining_time": "2:09:21"}
-{"current_steps": 14, "total_steps": 90, "loss": 1.3799, "lr": 9.906276553136923e-05, "epoch": 0.46473029045643155, "percentage": 15.56, "elapsed_time": "0:23:11", "remaining_time": "2:05:52"}
-{"current_steps": 15, "total_steps": 90, "loss": 1.4019, "lr": 9.865224352899119e-05, "epoch": 0.4979253112033195, "percentage": 16.67, "elapsed_time": "0:24:40", "remaining_time": "2:03:24"}
-{"current_steps": 16, "total_steps": 90, "loss": 1.4167, "lr": 9.816854393079403e-05, "epoch": 0.5311203319502075, "percentage": 17.78, "elapsed_time": "0:26:05", "remaining_time": "2:00:38"}
-{"current_steps": 17, "total_steps": 90, "loss": 1.3623, "lr": 9.761239426692077e-05, "epoch": 0.5643153526970954, "percentage": 18.89, "elapsed_time": "0:27:51", "remaining_time": "1:59:35"}
-{"current_steps": 18, "total_steps": 90, "loss": 1.394, "lr": 9.698463103929542e-05, "epoch": 0.5975103734439834, "percentage": 20.0, "elapsed_time": "0:29:24", "remaining_time": "1:57:39"}
-{"current_steps": 19, "total_steps": 90, "loss": 1.3858, "lr": 9.628619846344454e-05, "epoch": 0.6307053941908713, "percentage": 21.11, "elapsed_time": "0:31:01", "remaining_time": "1:55:57"}
-{"current_steps": 20, "total_steps": 90, "loss": 1.4395, "lr": 9.551814704830734e-05, "epoch": 0.6639004149377593, "percentage": 22.22, "elapsed_time": "0:32:30", "remaining_time": "1:53:45"}
-{"current_steps": 21, "total_steps": 90, "loss": 1.4441, "lr": 9.468163201617062e-05, "epoch": 0.6970954356846473, "percentage": 23.33, "elapsed_time": "0:34:11", "remaining_time": "1:52:20"}
-{"current_steps": 22, "total_steps": 90, "loss": 1.3777, "lr": 9.377791156510455e-05, "epoch": 0.7302904564315352, "percentage": 24.44, "elapsed_time": "0:35:42", "remaining_time": "1:50:23"}
-{"current_steps": 23, "total_steps": 90, "loss": 1.3492, "lr": 9.280834497651334e-05, "epoch": 0.7634854771784232, "percentage": 25.56, "elapsed_time": "0:37:08", "remaining_time": "1:48:11"}
-{"current_steps": 24, "total_steps": 90, "loss": 1.3356, "lr": 9.177439057064683e-05, "epoch": 0.7966804979253111, "percentage": 26.67, "elapsed_time": "0:38:40", "remaining_time": "1:46:21"}
-{"current_steps": 25, "total_steps": 90, "loss": 1.4073, "lr": 9.067760351314838e-05, "epoch": 0.8298755186721992, "percentage": 27.78, "elapsed_time": "0:40:08", "remaining_time": "1:44:21"}
-{"current_steps": 26, "total_steps": 90, "loss": 1.3822, "lr": 8.951963347593797e-05, "epoch": 0.8630705394190872, "percentage": 28.89, "elapsed_time": "0:41:29", "remaining_time": "1:42:08"}
-{"current_steps": 27, "total_steps": 90, "loss": 1.3754, "lr": 8.83022221559489e-05, "epoch": 0.8962655601659751, "percentage": 30.0, "elapsed_time": "0:42:51", "remaining_time": "1:40:00"}
-{"current_steps": 28, "total_steps": 90, "loss": 1.3959, "lr": 8.702720065545024e-05, "epoch": 0.9294605809128631, "percentage": 31.11, "elapsed_time": "0:44:14", "remaining_time": "1:37:57"}
-{"current_steps": 29, "total_steps": 90, "loss": 1.3844, "lr": 8.569648672789497e-05, "epoch": 0.9626556016597511, "percentage": 32.22, "elapsed_time": "0:45:51", "remaining_time": "1:36:28"}
-{"current_steps": 30, "total_steps": 90, "loss": 1.4104, "lr": 8.43120818934367e-05, "epoch": 0.995850622406639, "percentage": 33.33, "elapsed_time": "0:47:28", "remaining_time": "1:34:56"}
-{"current_steps": 31, "total_steps": 90, "loss": 2.4088, "lr": 8.28760684284532e-05, "epoch": 1.0290456431535269, "percentage": 34.44, "elapsed_time": "0:49:21", "remaining_time": "1:33:57"}
-{"current_steps": 32, "total_steps": 90, "loss": 1.3963, "lr": 8.139060623360493e-05, "epoch": 1.062240663900415, "percentage": 35.56, "elapsed_time": "0:51:21", "remaining_time": "1:33:04"}
-{"current_steps": 33, "total_steps": 90, "loss": 1.3139, "lr": 7.985792958513931e-05, "epoch": 1.095435684647303, "percentage": 36.67, "elapsed_time": "0:53:04", "remaining_time": "1:31:40"}
-{"current_steps": 34, "total_steps": 90, "loss": 1.2728, "lr": 7.828034377432693e-05, "epoch": 1.1286307053941909, "percentage": 37.78, "elapsed_time": "0:54:38", "remaining_time": "1:29:59"}
-{"current_steps": 35, "total_steps": 90, "loss": 1.3413, "lr": 7.666022164008457e-05, "epoch": 1.161825726141079, "percentage": 38.89, "elapsed_time": "0:56:10", "remaining_time": "1:28:17"}
-{"current_steps": 36, "total_steps": 90, "loss": 1.2752, "lr": 7.500000000000001e-05, "epoch": 1.1950207468879668, "percentage": 40.0, "elapsed_time": "0:57:16", "remaining_time": "1:25:54"}
-{"current_steps": 37, "total_steps": 90, "loss": 1.3097, "lr": 7.330217598512695e-05, "epoch": 1.2282157676348548, "percentage": 41.11, "elapsed_time": "0:58:45", "remaining_time": "1:24:10"}
-{"current_steps": 38, "total_steps": 90, "loss": 1.3755, "lr": 7.156930328406268e-05, "epoch": 1.2614107883817427, "percentage": 42.22, "elapsed_time": "1:00:32", "remaining_time": "1:22:51"}
-{"current_steps": 39, "total_steps": 90, "loss": 1.3168, "lr": 6.980398830195785e-05, "epoch": 1.2946058091286308, "percentage": 43.33, "elapsed_time": "1:02:11", "remaining_time": "1:21:20"}
-{"current_steps": 40, "total_steps": 90, "loss": 1.324, "lr": 6.800888624023553e-05, "epoch": 1.3278008298755186, "percentage": 44.44, "elapsed_time": "1:04:03", "remaining_time": "1:20:04"}
-{"current_steps": 41, "total_steps": 90, "loss": 1.3157, "lr": 6.618669710291606e-05, "epoch": 1.3609958506224067, "percentage": 45.56, "elapsed_time": "1:05:26", "remaining_time": "1:18:12"}
-{"current_steps": 42, "total_steps": 90, "loss": 1.3328, "lr": 6.434016163555452e-05, "epoch": 1.3941908713692945, "percentage": 46.67, "elapsed_time": "1:07:03", "remaining_time": "1:16:38"}
-{"current_steps": 43, "total_steps": 90, "loss": 1.3266, "lr": 6.247205720289907e-05, "epoch": 1.4273858921161826, "percentage": 47.78, "elapsed_time": "1:08:57", "remaining_time": "1:15:21"}
-{"current_steps": 44, "total_steps": 90, "loss": 1.2572, "lr": 6.058519361147055e-05, "epoch": 1.4605809128630705, "percentage": 48.89, "elapsed_time": "1:10:23", "remaining_time": "1:13:35"}
-{"current_steps": 45, "total_steps": 90, "loss": 1.2573, "lr": 5.868240888334653e-05, "epoch": 1.4937759336099585, "percentage": 50.0, "elapsed_time": "1:11:32", "remaining_time": "1:11:32"}
-{"current_steps": 46, "total_steps": 90, "loss": 1.3751, "lr": 5.6766564987506566e-05, "epoch": 1.5269709543568464, "percentage": 51.11, "elapsed_time": "1:13:25", "remaining_time": "1:10:13"}
-{"current_steps": 47, "total_steps": 90, "loss": 1.2813, "lr": 5.484054353515896e-05, "epoch": 1.5601659751037344, "percentage": 52.22, "elapsed_time": "1:15:00", "remaining_time": "1:08:37"}
-{"current_steps": 48, "total_steps": 90, "loss": 1.3213, "lr": 5.290724144552379e-05, "epoch": 1.5933609958506225, "percentage": 53.33, "elapsed_time": "1:16:54", "remaining_time": "1:07:17"}
-{"current_steps": 49, "total_steps": 90, "loss": 1.2277, "lr": 5.096956658859122e-05, "epoch": 1.6265560165975104, "percentage": 54.44, "elapsed_time": "1:18:42", "remaining_time": "1:05:51"}
-{"current_steps": 50, "total_steps": 90, "loss": 1.3067, "lr": 4.903043341140879e-05, "epoch": 1.6597510373443982, "percentage": 55.56, "elapsed_time": "1:19:59", "remaining_time": "1:03:59"}
-{"current_steps": 51, "total_steps": 90, "loss": 1.3133, "lr": 4.709275855447621e-05, "epoch": 1.6929460580912863, "percentage": 56.67, "elapsed_time": "1:21:50", "remaining_time": "1:02:35"}
-{"current_steps": 52, "total_steps": 90, "loss": 1.2421, "lr": 4.515945646484105e-05, "epoch": 1.7261410788381744, "percentage": 57.78, "elapsed_time": "1:23:10", "remaining_time": "1:00:46"}
-{"current_steps": 53, "total_steps": 90, "loss": 1.315, "lr": 4.323343501249346e-05, "epoch": 1.7593360995850622, "percentage": 58.89, "elapsed_time": "1:25:13", "remaining_time": "0:59:29"}
-{"current_steps": 54, "total_steps": 90, "loss": 1.3178, "lr": 4.131759111665349e-05, "epoch": 1.79253112033195, "percentage": 60.0, "elapsed_time": "1:26:30", "remaining_time": "0:57:40"}
-{"current_steps": 55, "total_steps": 90, "loss": 1.2787, "lr": 3.941480638852948e-05, "epoch": 1.8257261410788381, "percentage": 61.11, "elapsed_time": "1:28:22", "remaining_time": "0:56:14"}
-{"current_steps": 56, "total_steps": 90, "loss": 1.2566, "lr": 3.752794279710094e-05, "epoch": 1.8589211618257262, "percentage": 62.22, "elapsed_time": "1:29:51", "remaining_time": "0:54:33"}
-{"current_steps": 57, "total_steps": 90, "loss": 1.3124, "lr": 3.5659838364445505e-05, "epoch": 1.892116182572614, "percentage": 63.33, "elapsed_time": "1:31:40", "remaining_time": "0:53:04"}
-{"current_steps": 58, "total_steps": 90, "loss": 1.2544, "lr": 3.381330289708396e-05, "epoch": 1.9253112033195021, "percentage": 64.44, "elapsed_time": "1:33:27", "remaining_time": "0:51:33"}
-{"current_steps": 59, "total_steps": 90, "loss": 1.2155, "lr": 3.199111375976449e-05, "epoch": 1.9585062240663902, "percentage": 65.56, "elapsed_time": "1:34:54", "remaining_time": "0:49:51"}
-{"current_steps": 60, "total_steps": 90, "loss": 1.2942, "lr": 3.019601169804216e-05, "epoch": 1.991701244813278, "percentage": 66.67, "elapsed_time": "1:36:09", "remaining_time": "0:48:04"}
-{"current_steps": 61, "total_steps": 90, "loss": 2.4636, "lr": 2.8430696715937337e-05, "epoch": 2.024896265560166, "percentage": 67.78, "elapsed_time": "1:37:29", "remaining_time": "0:46:20"}
-{"current_steps": 62, "total_steps": 90, "loss": 1.2014, "lr": 2.6697824014873075e-05, "epoch": 2.0580912863070537, "percentage": 68.89, "elapsed_time": "1:38:58", "remaining_time": "0:44:42"}
-{"current_steps": 63, "total_steps": 90, "loss": 1.2215, "lr": 2.500000000000001e-05, "epoch": 2.091286307053942, "percentage": 70.0, "elapsed_time": "1:41:05", "remaining_time": "0:43:19"}
-{"current_steps": 64, "total_steps": 90, "loss": 1.2699, "lr": 2.333977835991545e-05, "epoch": 2.12448132780083, "percentage": 71.11, "elapsed_time": "1:42:47", "remaining_time": "0:41:45"}
-{"current_steps": 65, "total_steps": 90, "loss": 1.0933, "lr": 2.171965622567308e-05, "epoch": 2.1576763485477177, "percentage": 72.22, "elapsed_time": "1:44:04", "remaining_time": "0:40:01"}
-{"current_steps": 66, "total_steps": 90, "loss": 1.4325, "lr": 2.0142070414860704e-05, "epoch": 2.190871369294606, "percentage": 73.33, "elapsed_time": "1:46:14", "remaining_time": "0:38:38"}
-{"current_steps": 67, "total_steps": 90, "loss": 1.1395, "lr": 1.8609393766395085e-05, "epoch": 2.224066390041494, "percentage": 74.44, "elapsed_time": "1:47:51", "remaining_time": "0:37:01"}
-{"current_steps": 68, "total_steps": 90, "loss": 1.3358, "lr": 1.7123931571546827e-05, "epoch": 2.2572614107883817, "percentage": 75.56, "elapsed_time": "1:49:14", "remaining_time": "0:35:20"}
-{"current_steps": 69, "total_steps": 90, "loss": 1.1908, "lr": 1.5687918106563326e-05, "epoch": 2.2904564315352696, "percentage": 76.67, "elapsed_time": "1:50:23", "remaining_time": "0:33:35"}
-{"current_steps": 70, "total_steps": 90, "loss": 1.2044, "lr": 1.4303513272105057e-05, "epoch": 2.323651452282158, "percentage": 77.78, "elapsed_time": "1:51:59", "remaining_time": "0:31:59"}
-{"current_steps": 71, "total_steps": 90, "loss": 1.2635, "lr": 1.297279934454978e-05, "epoch": 2.3568464730290457, "percentage": 78.89, "elapsed_time": "1:53:45", "remaining_time": "0:30:26"}
-{"current_steps": 72, "total_steps": 90, "loss": 1.1812, "lr": 1.1697777844051105e-05, "epoch": 2.3900414937759336, "percentage": 80.0, "elapsed_time": "1:55:14", "remaining_time": "0:28:48"}
-{"current_steps": 73, "total_steps": 90, "loss": 1.3095, "lr": 1.0480366524062042e-05, "epoch": 2.4232365145228214, "percentage": 81.11, "elapsed_time": "1:56:54", "remaining_time": "0:27:13"}
-{"current_steps": 74, "total_steps": 90, "loss": 1.1945, "lr": 9.322396486851626e-06, "epoch": 2.4564315352697097, "percentage": 82.22, "elapsed_time": "1:58:23", "remaining_time": "0:25:35"}
-{"current_steps": 75, "total_steps": 90, "loss": 1.2426, "lr": 8.225609429353187e-06, "epoch": 2.4896265560165975, "percentage": 83.33, "elapsed_time": "2:00:04", "remaining_time": "0:24:00"}
-{"current_steps": 76, "total_steps": 90, "loss": 1.2592, "lr": 7.191655023486682e-06, "epoch": 2.5228215767634854, "percentage": 84.44, "elapsed_time": "2:01:39", "remaining_time": "0:22:24"}
-{"current_steps": 77, "total_steps": 90, "loss": 1.2888, "lr": 6.222088434895462e-06, "epoch": 2.5560165975103732, "percentage": 85.56, "elapsed_time": "2:03:28", "remaining_time": "0:20:50"}
-{"current_steps": 78, "total_steps": 90, "loss": 1.1362, "lr": 5.318367983829392e-06, "epoch": 2.5892116182572615, "percentage": 86.67, "elapsed_time": "2:05:27", "remaining_time": "0:19:18"}
-{"current_steps": 79, "total_steps": 90, "loss": 1.3568, "lr": 4.4818529516926726e-06, "epoch": 2.6224066390041494, "percentage": 87.78, "elapsed_time": "2:06:54", "remaining_time": "0:17:40"}
-{"current_steps": 80, "total_steps": 90, "loss": 1.2015, "lr": 3.7138015365554833e-06, "epoch": 2.6556016597510372, "percentage": 88.89, "elapsed_time": "2:08:22", "remaining_time": "0:16:02"}
-{"current_steps": 81, "total_steps": 90, "loss": 1.2846, "lr": 3.0153689607045845e-06, "epoch": 2.6887966804979255, "percentage": 90.0, "elapsed_time": "2:09:55", "remaining_time": "0:14:26"}
-{"current_steps": 82, "total_steps": 90, "loss": 1.193, "lr": 2.3876057330792346e-06, "epoch": 2.7219917012448134, "percentage": 91.11, "elapsed_time": "2:11:56", "remaining_time": "0:12:52"}
-{"current_steps": 83, "total_steps": 90, "loss": 1.3376, "lr": 1.8314560692059835e-06, "epoch": 2.7551867219917012, "percentage": 92.22, "elapsed_time": "2:13:42", "remaining_time": "0:11:16"}
-{"current_steps": 84, "total_steps": 90, "loss": 1.2094, "lr": 1.3477564710088098e-06, "epoch": 2.788381742738589, "percentage": 93.33, "elapsed_time": "2:15:29", "remaining_time": "0:09:40"}
-{"current_steps": 85, "total_steps": 90, "loss": 1.2275, "lr": 9.372344686307655e-07, "epoch": 2.821576763485477, "percentage": 94.44, "elapsed_time": "2:17:02", "remaining_time": "0:08:03"}
-{"current_steps": 86, "total_steps": 90, "loss": 1.2576, "lr": 6.005075261595494e-07, "epoch": 2.854771784232365, "percentage": 95.56, "elapsed_time": "2:18:41", "remaining_time": "0:06:27"}
-{"current_steps": 87, "total_steps": 90, "loss": 1.254, "lr": 3.380821129028489e-07, "epoch": 2.887966804979253, "percentage": 96.67, "elapsed_time": "2:20:24", "remaining_time": "0:04:50"}
-{"current_steps": 88, "total_steps": 90, "loss": 1.327, "lr": 1.503529416103988e-07, "epoch": 2.921161825726141, "percentage": 97.78, "elapsed_time": "2:22:11", "remaining_time": "0:03:13"}
-{"current_steps": 89, "total_steps": 90, "loss": 1.1686, "lr": 3.760237478849793e-08, "epoch": 2.954356846473029, "percentage": 98.89, "elapsed_time": "2:23:42", "remaining_time": "0:01:36"}
-{"current_steps": 90, "total_steps": 90, "loss": 1.1824, "lr": 0.0, "epoch": 2.987551867219917, "percentage": 100.0, "elapsed_time": "2:25:23", "remaining_time": "0:00:00"}
-{"current_steps": 90, "total_steps": 90, "epoch": 2.987551867219917, "percentage": 100.0, "elapsed_time": "2:25:56", "remaining_time": "0:00:00"}

llama_factory/v3-20241225-1_5/trainer_state.json

@@ -1,672 +0,0 @@
-{
-  "best_metric": null,
-  "best_model_checkpoint": null,
-  "epoch": 2.987551867219917,
-  "eval_steps": 500,
-  "global_step": 90,
-  "is_hyper_param_search": false,
-  "is_local_process_zero": true,
-  "is_world_process_zero": true,
-  "log_history": [
-    {
-      "epoch": 0.03319502074688797,
-      "grad_norm": 1.4870719909667969,
-      "learning_rate": 1.1111111111111112e-05,
-      "loss": 1.6367,
-      "step": 1
-    },
-    {
-      "epoch": 0.06639004149377593,
-      "grad_norm": 1.4453295469284058,
-      "learning_rate": 2.2222222222222223e-05,
-      "loss": 1.5947,
-      "step": 2
-    },
-    {
-      "epoch": 0.0995850622406639,
-      "grad_norm": 1.264766812324524,
-      "learning_rate": 3.3333333333333335e-05,
-      "loss": 1.6033,
-      "step": 3
-    },
-    {
-      "epoch": 0.13278008298755187,
-      "grad_norm": 0.6283169984817505,
-      "learning_rate": 4.4444444444444447e-05,
-      "loss": 1.4665,
-      "step": 4
-    },
-    {
-      "epoch": 0.16597510373443983,
-      "grad_norm": 0.36272674798965454,
-      "learning_rate": 5.555555555555556e-05,
-      "loss": 1.4843,
-      "step": 5
-    },
-    {
-      "epoch": 0.1991701244813278,
-      "grad_norm": 0.47409433126449585,
-      "learning_rate": 6.666666666666667e-05,
-      "loss": 1.5095,
-      "step": 6
-    },
-    {
-      "epoch": 0.23236514522821577,
-      "grad_norm": 0.5489184260368347,
-      "learning_rate": 7.777777777777778e-05,
-      "loss": 1.5704,
-      "step": 7
-    },
-    {
-      "epoch": 0.26556016597510373,
-      "grad_norm": 0.48931849002838135,
-      "learning_rate": 8.888888888888889e-05,
-      "loss": 1.4808,
-      "step": 8
-    },
-    {
-      "epoch": 0.2987551867219917,
-      "grad_norm": 0.310524046421051,
-      "learning_rate": 0.0001,
-      "loss": 1.4379,
-      "step": 9
-    },
-    {
-      "epoch": 0.33195020746887965,
-      "grad_norm": 0.2177194356918335,
-      "learning_rate": 9.996239762521151e-05,
-      "loss": 1.3778,
-      "step": 10
-    },
-    {
-      "epoch": 0.3651452282157676,
-      "grad_norm": 0.26418522000312805,
-      "learning_rate": 9.98496470583896e-05,
-      "loss": 1.4325,
-      "step": 11
-    },
-    {
-      "epoch": 0.3983402489626556,
-      "grad_norm": 0.28866058588027954,
-      "learning_rate": 9.966191788709716e-05,
-      "loss": 1.4703,
-      "step": 12
-    },
-    {
-      "epoch": 0.4315352697095436,
-      "grad_norm": 0.2462608516216278,
-      "learning_rate": 9.939949247384046e-05,
-      "loss": 1.4177,
-      "step": 13
-    },
-    {
-      "epoch": 0.46473029045643155,
-      "grad_norm": 0.179249107837677,
-      "learning_rate": 9.906276553136923e-05,
-      "loss": 1.3799,
-      "step": 14
-    },
-    {
-      "epoch": 0.4979253112033195,
-      "grad_norm": 0.220245361328125,
-      "learning_rate": 9.865224352899119e-05,
-      "loss": 1.4019,
-      "step": 15
-    },
-    {
-      "epoch": 0.5311203319502075,
-      "grad_norm": 0.23056794703006744,
-      "learning_rate": 9.816854393079403e-05,
-      "loss": 1.4167,
-      "step": 16
-    },
-    {
-      "epoch": 0.5643153526970954,
-      "grad_norm": 0.19931739568710327,
-      "learning_rate": 9.761239426692077e-05,
-      "loss": 1.3623,
-      "step": 17
-    },
-    {
-      "epoch": 0.5975103734439834,
-      "grad_norm": 0.1767573207616806,
-      "learning_rate": 9.698463103929542e-05,
-      "loss": 1.394,
-      "step": 18
-    },
-    {
-      "epoch": 0.6307053941908713,
-      "grad_norm": 0.17220574617385864,
-      "learning_rate": 9.628619846344454e-05,
-      "loss": 1.3858,
-      "step": 19
-    },
-    {
-      "epoch": 0.6639004149377593,
-      "grad_norm": 0.20925532281398773,
-      "learning_rate": 9.551814704830734e-05,
-      "loss": 1.4395,
-      "step": 20
-    },
-    {
-      "epoch": 0.6970954356846473,
-      "grad_norm": 0.19617941975593567,
-      "learning_rate": 9.468163201617062e-05,
-      "loss": 1.4441,
-      "step": 21
-    },
-    {
-      "epoch": 0.7302904564315352,
-      "grad_norm": 0.262824684381485,
-      "learning_rate": 9.377791156510455e-05,
-      "loss": 1.3777,
-      "step": 22
-    },
-    {
-      "epoch": 0.7634854771784232,
-      "grad_norm": 0.16262340545654297,
-      "learning_rate": 9.280834497651334e-05,
-      "loss": 1.3492,
-      "step": 23
-    },
-    {
-      "epoch": 0.7966804979253111,
-      "grad_norm": 0.16318321228027344,
-      "learning_rate": 9.177439057064683e-05,
-      "loss": 1.3356,
-      "step": 24
-    },
-    {
-      "epoch": 0.8298755186721992,
-      "grad_norm": 0.16352789103984833,
-      "learning_rate": 9.067760351314838e-05,
-      "loss": 1.4073,
-      "step": 25
-    },
-    {
-      "epoch": 0.8630705394190872,
-      "grad_norm": 0.1612463891506195,
-      "learning_rate": 8.951963347593797e-05,
-      "loss": 1.3822,
-      "step": 26
-    },
-    {
-      "epoch": 0.8962655601659751,
-      "grad_norm": 0.15770702064037323,
-      "learning_rate": 8.83022221559489e-05,
-      "loss": 1.3754,
-      "step": 27
-    },
-    {
-      "epoch": 0.9294605809128631,
-      "grad_norm": 0.15616007149219513,
-      "learning_rate": 8.702720065545024e-05,
-      "loss": 1.3959,
-      "step": 28
-    },
-    {
-      "epoch": 0.9626556016597511,
-      "grad_norm": 0.1599518060684204,
-      "learning_rate": 8.569648672789497e-05,
-      "loss": 1.3844,
-      "step": 29
-    },
-    {
-      "epoch": 0.995850622406639,
-      "grad_norm": 0.16726341843605042,
-      "learning_rate": 8.43120818934367e-05,
-      "loss": 1.4104,
-      "step": 30
-    },
-    {
-      "epoch": 1.0290456431535269,
-      "grad_norm": 0.41314148902893066,
-      "learning_rate": 8.28760684284532e-05,
-      "loss": 2.4088,
-      "step": 31
-    },
-    {
-      "epoch": 1.062240663900415,
-      "grad_norm": 0.15744566917419434,
-      "learning_rate": 8.139060623360493e-05,
-      "loss": 1.3963,
-      "step": 32
-    },
-    {
-      "epoch": 1.095435684647303,
-      "grad_norm": 0.16752980649471283,
-      "learning_rate": 7.985792958513931e-05,
-      "loss": 1.3139,
-      "step": 33
-    },
-    {
-      "epoch": 1.1286307053941909,
-      "grad_norm": 0.16712723672389984,
-      "learning_rate": 7.828034377432693e-05,
-      "loss": 1.2728,
-      "step": 34
-    },
-    {
-      "epoch": 1.161825726141079,
-      "grad_norm": 0.16366273164749146,
-      "learning_rate": 7.666022164008457e-05,
-      "loss": 1.3413,
-      "step": 35
-    },
-    {
-      "epoch": 1.1950207468879668,
-      "grad_norm": 0.14664730429649353,
-      "learning_rate": 7.500000000000001e-05,
-      "loss": 1.2752,
-      "step": 36
-    },
-    {
-      "epoch": 1.2282157676348548,
-      "grad_norm": 0.16132752597332,
-      "learning_rate": 7.330217598512695e-05,
-      "loss": 1.3097,
-      "step": 37
-    },
-    {
-      "epoch": 1.2614107883817427,
-      "grad_norm": 0.16922235488891602,
-      "learning_rate": 7.156930328406268e-05,
-      "loss": 1.3755,
-      "step": 38
-    },
-    {
-      "epoch": 1.2946058091286308,
-      "grad_norm": 0.16416330635547638,
-      "learning_rate": 6.980398830195785e-05,
-      "loss": 1.3168,
-      "step": 39
-    },
-    {
-      "epoch": 1.3278008298755186,
-      "grad_norm": 0.16735291481018066,
-      "learning_rate": 6.800888624023553e-05,
-      "loss": 1.324,
-      "step": 40
-    },
-    {
-      "epoch": 1.3609958506224067,
-      "grad_norm": 0.16032138466835022,
-      "learning_rate": 6.618669710291606e-05,
-      "loss": 1.3157,
-      "step": 41
-    },
-    {
-      "epoch": 1.3941908713692945,
-      "grad_norm": 0.16943234205245972,
-      "learning_rate": 6.434016163555452e-05,
-      "loss": 1.3328,
-      "step": 42
-    },
-    {
-      "epoch": 1.4273858921161826,
-      "grad_norm": 0.16917924582958221,
-      "learning_rate": 6.247205720289907e-05,
-      "loss": 1.3266,
-      "step": 43
-    },
-    {
-      "epoch": 1.4605809128630705,
-      "grad_norm": 0.16203759610652924,
-      "learning_rate": 6.058519361147055e-05,
-      "loss": 1.2572,
-      "step": 44
-    },
-    {
-      "epoch": 1.4937759336099585,
-      "grad_norm": 0.16321130096912384,
-      "learning_rate": 5.868240888334653e-05,
-      "loss": 1.2573,
-      "step": 45
-    },
-    {
-      "epoch": 1.5269709543568464,
-      "grad_norm": 0.16734814643859863,
-      "learning_rate": 5.6766564987506566e-05,
-      "loss": 1.3751,
-      "step": 46
-    },
-    {
-      "epoch": 1.5601659751037344,
-      "grad_norm": 0.15714752674102783,
-      "learning_rate": 5.484054353515896e-05,
-      "loss": 1.2813,
-      "step": 47
-    },
-    {
-      "epoch": 1.5933609958506225,
-      "grad_norm": 0.16355480253696442,
-      "learning_rate": 5.290724144552379e-05,
-      "loss": 1.3213,
-      "step": 48
-    },
-    {
-      "epoch": 1.6265560165975104,
-      "grad_norm": 0.1648782640695572,
-      "learning_rate": 5.096956658859122e-05,
-      "loss": 1.2277,
-      "step": 49
-    },
-    {
-      "epoch": 1.6597510373443982,
-      "grad_norm": 0.17283305525779724,
-      "learning_rate": 4.903043341140879e-05,
-      "loss": 1.3067,
-      "step": 50
-    },
-    {
-      "epoch": 1.6929460580912863,
-      "grad_norm": 0.16521140933036804,
-      "learning_rate": 4.709275855447621e-05,
-      "loss": 1.3133,
-      "step": 51
-    },
-    {
-      "epoch": 1.7261410788381744,
-      "grad_norm": 0.1587812900543213,
-      "learning_rate": 4.515945646484105e-05,
-      "loss": 1.2421,
-      "step": 52
-    },
-    {
-      "epoch": 1.7593360995850622,
-      "grad_norm": 0.16004517674446106,
-      "learning_rate": 4.323343501249346e-05,
-      "loss": 1.315,
-      "step": 53
-    },
-    {
-      "epoch": 1.79253112033195,
-      "grad_norm": 0.16503478586673737,
-      "learning_rate": 4.131759111665349e-05,
-      "loss": 1.3178,
-      "step": 54
-    },
-    {
-      "epoch": 1.8257261410788381,
-      "grad_norm": 0.15602463483810425,
-      "learning_rate": 3.941480638852948e-05,
-      "loss": 1.2787,
-      "step": 55
-    },
-    {
-      "epoch": 1.8589211618257262,
-      "grad_norm": 0.17025011777877808,
-      "learning_rate": 3.752794279710094e-05,
-      "loss": 1.2566,
-      "step": 56
-    },
-    {
-      "epoch": 1.892116182572614,
-      "grad_norm": 0.1677195280790329,
-      "learning_rate": 3.5659838364445505e-05,
-      "loss": 1.3124,
-      "step": 57
-    },
-    {
-      "epoch": 1.9253112033195021,
-      "grad_norm": 0.16267016530036926,
-      "learning_rate": 3.381330289708396e-05,
-      "loss": 1.2544,
-      "step": 58
-    },
-    {
-      "epoch": 1.9585062240663902,
-      "grad_norm": 0.16983366012573242,
-      "learning_rate": 3.199111375976449e-05,
-      "loss": 1.2155,
-      "step": 59
-    },
-    {
-      "epoch": 1.991701244813278,
-      "grad_norm": 0.16994623839855194,
-      "learning_rate": 3.019601169804216e-05,
-      "loss": 1.2942,
-      "step": 60
-    },
-    {
-      "epoch": 2.024896265560166,
-      "grad_norm": 0.4933292269706726,
-      "learning_rate": 2.8430696715937337e-05,
-      "loss": 2.4636,
-      "step": 61
-    },
-    {
-      "epoch": 2.0580912863070537,
-      "grad_norm": 0.16370011866092682,
-      "learning_rate": 2.6697824014873075e-05,
-      "loss": 1.2014,
-      "step": 62
-    },
-    {
-      "epoch": 2.091286307053942,
-      "grad_norm": 0.16958041489124298,
-      "learning_rate": 2.500000000000001e-05,
-      "loss": 1.2215,
-      "step": 63
-    },
-    {
-      "epoch": 2.12448132780083,
-      "grad_norm": 0.17015086114406586,
-      "learning_rate": 2.333977835991545e-05,
-      "loss": 1.2699,
-      "step": 64
-    },
-    {
-      "epoch": 2.1576763485477177,
-      "grad_norm": 0.15585944056510925,
-      "learning_rate": 2.171965622567308e-05,
-      "loss": 1.0933,
-      "step": 65
-    },
-    {
-      "epoch": 2.190871369294606,
-      "grad_norm": 0.17827057838439941,
-      "learning_rate": 2.0142070414860704e-05,
-      "loss": 1.4325,
-      "step": 66
-    },
-    {
-      "epoch": 2.224066390041494,
-      "grad_norm": 0.1636652946472168,
-      "learning_rate": 1.8609393766395085e-05,
-      "loss": 1.1395,
-      "step": 67
-    },
-    {
-      "epoch": 2.2572614107883817,
-      "grad_norm": 0.1846497654914856,
-      "learning_rate": 1.7123931571546827e-05,
-      "loss": 1.3358,
-      "step": 68
-    },
-    {
-      "epoch": 2.2904564315352696,
-      "grad_norm": 0.17364221811294556,
-      "learning_rate": 1.5687918106563326e-05,
-      "loss": 1.1908,
-      "step": 69
-    },
-    {
-      "epoch": 2.323651452282158,
-      "grad_norm": 0.17447641491889954,
-      "learning_rate": 1.4303513272105057e-05,
-      "loss": 1.2044,
-      "step": 70
-    },
-    {
-      "epoch": 2.3568464730290457,
-      "grad_norm": 0.18103092908859253,
-      "learning_rate": 1.297279934454978e-05,
-      "loss": 1.2635,
-      "step": 71
-    },
-    {
-      "epoch": 2.3900414937759336,
-      "grad_norm": 0.16945283114910126,
-      "learning_rate": 1.1697777844051105e-05,
-      "loss": 1.1812,
-      "step": 72
-    },
-    {
-      "epoch": 2.4232365145228214,
-      "grad_norm": 0.17960192263126373,
-      "learning_rate": 1.0480366524062042e-05,
-      "loss": 1.3095,
-      "step": 73
-    },
-    {
-      "epoch": 2.4564315352697097,
-      "grad_norm": 0.17297744750976562,
-      "learning_rate": 9.322396486851626e-06,
-      "loss": 1.1945,
-      "step": 74
-    },
-    {
-      "epoch": 2.4896265560165975,
-      "grad_norm": 0.170962855219841,
-      "learning_rate": 8.225609429353187e-06,
-      "loss": 1.2426,
-      "step": 75
-    },
-    {
-      "epoch": 2.5228215767634854,
-      "grad_norm": 0.17505840957164764,
-      "learning_rate": 7.191655023486682e-06,
-      "loss": 1.2592,
-      "step": 76
-    },
-    {
-      "epoch": 2.5560165975103732,
-      "grad_norm": 0.17825698852539062,
-      "learning_rate": 6.222088434895462e-06,
-      "loss": 1.2888,
-      "step": 77
-    },
-    {
-      "epoch": 2.5892116182572615,
-      "grad_norm": 0.16860072314739227,
-      "learning_rate": 5.318367983829392e-06,
-      "loss": 1.1362,
-      "step": 78
-    },
-    {
-      "epoch": 2.6224066390041494,
-      "grad_norm": 0.19669616222381592,
-      "learning_rate": 4.4818529516926726e-06,
-      "loss": 1.3568,
-      "step": 79
-    },
-    {
-      "epoch": 2.6556016597510372,
-      "grad_norm": 0.18090999126434326,
-      "learning_rate": 3.7138015365554833e-06,
-      "loss": 1.2015,
-      "step": 80
-    },
-    {
-      "epoch": 2.6887966804979255,
-      "grad_norm": 0.18253636360168457,
-      "learning_rate": 3.0153689607045845e-06,
-      "loss": 1.2846,
-      "step": 81
-    },
-    {
-      "epoch": 2.7219917012448134,
-      "grad_norm": 0.16775169968605042,
-      "learning_rate": 2.3876057330792346e-06,
-      "loss": 1.193,
-      "step": 82
-    },
-    {
-      "epoch": 2.7551867219917012,
-      "grad_norm": 0.17936046421527863,
-      "learning_rate": 1.8314560692059835e-06,
-      "loss": 1.3376,
-      "step": 83
-    },
-    {
-      "epoch": 2.788381742738589,
-      "grad_norm": 0.17388027906417847,
-      "learning_rate": 1.3477564710088098e-06,
-      "loss": 1.2094,
-      "step": 84
-    },
-    {
-      "epoch": 2.821576763485477,
-      "grad_norm": 0.17538011074066162,
-      "learning_rate": 9.372344686307655e-07,
-      "loss": 1.2275,
-      "step": 85
-    },
-    {
-      "epoch": 2.854771784232365,
-      "grad_norm": 0.1731182485818863,
-      "learning_rate": 6.005075261595494e-07,
-      "loss": 1.2576,
-      "step": 86
-    },
-    {
-      "epoch": 2.887966804979253,
-      "grad_norm": 0.17777325212955475,
-      "learning_rate": 3.380821129028489e-07,
-      "loss": 1.254,
-      "step": 87
-    },
-    {
-      "epoch": 2.921161825726141,
-      "grad_norm": 0.1849503070116043,
-      "learning_rate": 1.503529416103988e-07,
-      "loss": 1.327,
-      "step": 88
-    },
-    {
-      "epoch": 2.954356846473029,
-      "grad_norm": 0.16624118387699127,
-      "learning_rate": 3.760237478849793e-08,
-      "loss": 1.1686,
-      "step": 89
-    },
-    {
-      "epoch": 2.987551867219917,
-      "grad_norm": 0.1639927178621292,
-      "learning_rate": 0.0,
-      "loss": 1.1824,
-      "step": 90
-    },
-    {
-      "epoch": 2.987551867219917,
-      "step": 90,
-      "total_flos": 1.9253998183608484e+18,
-      "train_loss": 1.3520980450842115,
-      "train_runtime": 8756.8894,
-      "train_samples_per_second": 0.659,
-      "train_steps_per_second": 0.01
-    }
-  ],
-  "logging_steps": 1,
-  "max_steps": 90,
-  "num_input_tokens_seen": 0,
-  "num_train_epochs": 3,
-  "save_steps": 200,
-  "stateful_callbacks": {
-    "TrainerControl": {
-      "args": {
-        "should_epoch_stop": false,
-        "should_evaluate": false,
-        "should_log": false,
-        "should_save": true,
-        "should_training_stop": true
-      },
-      "attributes": {}
-    }
-  },
-  "total_flos": 1.9253998183608484e+18,
-  "train_batch_size": 2,
-  "trial_name": null,
-  "trial_params": null
-}

llama_factory/v3-20241225-pos/README.md

@@ -1,62 +0,0 @@
----
-library_name: peft
-license: other
-base_model: /disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct
-tags:
-- llama-factory
-- lora
-- generated_from_trainer
-model-index:
-- name: v3-20241225-pos
-  results: []
----
-
-<!-- This model card has been generated automatically according to the information the Trainer had access to. You
-should probably proofread and complete it, then remove this comment. -->
-
-# v3-20241225-pos
-
-This model is a fine-tuned version of [/disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct](https://huggingface.co//disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct) on the M^2RAR-20241225-LLM-Pos dataset.
-
-## Model description
-
-More information needed
-
-## Intended uses & limitations
-
-More information needed
-
-## Training and evaluation data
-
-More information needed
-
-## Training procedure
-
-### Training hyperparameters
-
-The following hyperparameters were used during training:
-- learning_rate: 0.0001
-- train_batch_size: 2
-- eval_batch_size: 8
-- seed: 42
-- distributed_type: multi-GPU
-- num_devices: 2
-- gradient_accumulation_steps: 16
-- total_train_batch_size: 64
-- total_eval_batch_size: 16
-- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
-- lr_scheduler_type: cosine
-- lr_scheduler_warmup_ratio: 0.1
-- num_epochs: 3.0
-
-### Training results
-
-
-
-### Framework versions
-
-- PEFT 0.12.0
-- Transformers 4.46.1
-- Pytorch 2.3.1+cu121
-- Datasets 3.1.0
-- Tokenizers 0.20.3

llama_factory/v3-20241225-pos/adapter_config.json

@@ -1,34 +0,0 @@
-{
-  "alpha_pattern": {},
-  "auto_mapping": null,
-  "base_model_name_or_path": "/disk/maziao/model-zoo/hf-repo/models--Qwen--Qwen2.5-7B-Instruct",
-  "bias": "none",
-  "fan_in_fan_out": false,
-  "inference_mode": true,
-  "init_lora_weights": true,
-  "layer_replication": null,
-  "layers_pattern": null,
-  "layers_to_transform": null,
-  "loftq_config": {},
-  "lora_alpha": 256,
-  "lora_dropout": 0.0,
-  "megatron_config": null,
-  "megatron_core": "megatron.core",
-  "modules_to_save": null,
-  "peft_type": "LORA",
-  "r": 128,
-  "rank_pattern": {},
-  "revision": null,
-  "target_modules": [
-    "o_proj",
-    "down_proj",
-    "q_proj",
-    "v_proj",
-    "gate_proj",
-    "up_proj",
-    "k_proj"
-  ],
-  "task_type": "CAUSAL_LM",
-  "use_dora": false,
-  "use_rslora": false
-}