Model save

README.md

@@ -1,5 +1,5 @@
 ---
-base_model: Qwen/Qwen2.5-0.5B
+base_model: Qwen/Qwen2.5-Math-7B
 library_name: transformers
 model_name: Qwen-2.5-7B-Simple-RL
 tags:
@@ -11,7 +11,7 @@ licence: license
 
 # Model Card for Qwen-2.5-7B-Simple-RL
 
-This model is a fine-tuned version of [Qwen/Qwen2.5-0.5B](https://huggingface.co/Qwen/Qwen2.5-0.5B).
+This model is a fine-tuned version of [Qwen/Qwen2.5-Math-7B](https://huggingface.co/Qwen/Qwen2.5-Math-7B).
 It has been trained using [TRL](https://github.com/huggingface/trl).
 
 ## Quick start

adapter_config.json

@@ -0,0 +1,37 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen2.5-Math-7B",
+  "bias": "none",
+  "eva_config": null,
+  "exclude_modules": null,
+  "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": 16,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "up_proj",
+    "o_proj",
+    "v_proj",
+    "gate_proj",
+    "k_proj",
+    "q_proj",
+    "down_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d1299b48026bfa4882d567b9562ebfc2772388a281c3593858b688dd0446f7c
+size 161533192

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": -0.002354813553392887,
-    "train_runtime": 733.4112,
+    "train_loss": 0.0008721697144210338,
+    "train_runtime": 2002.6709,
     "train_samples": 7500,
-    "train_samples_per_second": 1.309,
-    "train_steps_per_second": 0.041
+    "train_samples_per_second": 0.24,
+    "train_steps_per_second": 0.015
 }

config.json

@@ -23,7 +23,7 @@
   "tie_word_embeddings": true,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.49.0",
-  "use_cache": false,
+  "use_cache": true,
   "use_mrope": false,
   "use_sliding_window": false,
   "vocab_size": 151936

tokenizer_config.json

@@ -195,7 +195,7 @@
     "<|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 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 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",
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'Please reason step by step, and put your final answer within \\\\boxed{}.' }}\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\\nPlease reason step by step, and put your final answer within \\\\boxed{}.<|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": "<|endoftext|>",
   "errors": "replace",

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": -0.002354813553392887,
-    "train_runtime": 733.4112,
+    "train_loss": 0.0008721697144210338,
+    "train_runtime": 2002.6709,
     "train_samples": 7500,
-    "train_samples_per_second": 1.309,
-    "train_steps_per_second": 0.041
+    "train_samples_per_second": 0.24,
+    "train_steps_per_second": 0.015
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.064,
+  "epoch": 0.032,
   "eval_steps": 100,
   "global_step": 30,
   "is_hyper_param_search": false,
@@ -10,110 +10,110 @@
   "log_history": [
     {
       "clip_ratio": 0.0,
-      "completion_length": 576.28125,
-      "epoch": 0.0021333333333333334,
-      "grad_norm": 0.20850299298763275,
+      "completion_length": 246.8125,
+      "epoch": 0.0010666666666666667,
+      "grad_norm": 0.0,
       "kl": 0.0,
       "learning_rate": 1e-06,
-      "loss": -0.0284,
-      "reward": 0.03125,
-      "reward_std": 0.04419417306780815,
-      "rewards/accuracy_reward": 0.03125,
+      "loss": 0.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
       "rewards/format_reward": 0.0,
       "step": 1
     },
     {
       "clip_ratio": 0.0,
-      "completion_length": 571.2578125,
-      "epoch": 0.010666666666666666,
-      "grad_norm": 0.19187743961811066,
-      "kl": 0.0002953857183456421,
+      "completion_length": 228.984375,
+      "epoch": 0.005333333333333333,
+      "grad_norm": 0.0002897010708693415,
+      "kl": 0.0002434550688121817,
       "learning_rate": 2.959567305869736e-06,
-      "loss": -0.0078,
-      "reward": 0.0703125,
-      "reward_std": 0.05524271633476019,
-      "rewards/accuracy_reward": 0.0703125,
+      "loss": -0.0114,
+      "reward": 0.0625,
+      "reward_std": 0.0883883461356163,
+      "rewards/accuracy_reward": 0.0625,
       "rewards/format_reward": 0.0,
       "step": 5
     },
     {
       "clip_ratio": 0.0,
-      "completion_length": 595.875,
-      "epoch": 0.021333333333333333,
-      "grad_norm": 0.1668393611907959,
-      "kl": 0.0013158321380615234,
+      "completion_length": 234.275,
+      "epoch": 0.010666666666666666,
+      "grad_norm": 0.012982388027012348,
+      "kl": 0.00034311858398723417,
       "learning_rate": 2.529362456803101e-06,
-      "loss": -0.0104,
-      "reward": 0.1,
-      "reward_std": 0.07071067690849304,
-      "rewards/accuracy_reward": 0.1,
+      "loss": -0.0046,
+      "reward": 0.0625,
+      "reward_std": 0.01767766922712326,
+      "rewards/accuracy_reward": 0.0625,
       "rewards/format_reward": 0.0,
       "step": 10
     },
     {
       "clip_ratio": 0.0,
-      "completion_length": 580.5,
-      "epoch": 0.032,
-      "grad_norm": 0.5971764326095581,
-      "kl": 0.005566215515136719,
+      "completion_length": 239.0,
+      "epoch": 0.016,
+      "grad_norm": 0.00560641847550869,
+      "kl": 0.000386229114519665,
       "learning_rate": 1.7604722665003958e-06,
-      "loss": 0.0111,
-      "reward": 0.1,
-      "reward_std": 0.10606601536273956,
-      "rewards/accuracy_reward": 0.1,
+      "loss": 0.0031,
+      "reward": 0.1375,
+      "reward_std": 0.12374368458986282,
+      "rewards/accuracy_reward": 0.1375,
       "rewards/format_reward": 0.0,
       "step": 15
     },
     {
       "clip_ratio": 0.0,
-      "completion_length": 589.3125,
-      "epoch": 0.042666666666666665,
-      "grad_norm": 0.2944951355457306,
-      "kl": 0.00770721435546875,
+      "completion_length": 244.1125,
+      "epoch": 0.021333333333333333,
+      "grad_norm": 0.0002242429181933403,
+      "kl": 0.00041347902733832596,
       "learning_rate": 9.058803509412648e-07,
-      "loss": 0.0027,
-      "reward": 0.0875,
-      "reward_std": 0.07071067690849304,
-      "rewards/accuracy_reward": 0.0875,
+      "loss": 0.0022,
+      "reward": 0.05,
+      "reward_std": 0.03535533845424652,
+      "rewards/accuracy_reward": 0.05,
       "rewards/format_reward": 0.0,
       "step": 20
     },
     {
       "clip_ratio": 0.0,
-      "completion_length": 637.46875,
-      "epoch": 0.05333333333333334,
-      "grad_norm": 0.3306269943714142,
-      "kl": 0.005007171630859375,
+      "completion_length": 242.75,
+      "epoch": 0.02666666666666667,
+      "grad_norm": 0.016108760610222816,
+      "kl": 0.0003348791698954301,
       "learning_rate": 2.467682828805956e-07,
-      "loss": -0.0159,
-      "reward": 0.125,
-      "reward_std": 0.10606601536273956,
-      "rewards/accuracy_reward": 0.125,
+      "loss": 0.0156,
+      "reward": 0.1375,
+      "reward_std": 0.0883883461356163,
+      "rewards/accuracy_reward": 0.1375,
       "rewards/format_reward": 0.0,
       "step": 25
     },
     {
       "clip_ratio": 0.0,
-      "completion_length": 585.0625,
-      "epoch": 0.064,
-      "grad_norm": 0.4771800935268402,
-      "kl": 0.005538177490234375,
+      "completion_length": 234.625,
+      "epoch": 0.032,
+      "grad_norm": 0.0001786644133972004,
+      "kl": 0.0004210583254462108,
       "learning_rate": 0.0,
-      "loss": 0.0103,
-      "reward": 0.1,
+      "loss": -0.0019,
+      "reward": 0.0625,
       "reward_std": 0.05303300768136978,
-      "rewards/accuracy_reward": 0.1,
+      "rewards/accuracy_reward": 0.0625,
       "rewards/format_reward": 0.0,
       "step": 30
     },
     {
-      "epoch": 0.064,
+      "epoch": 0.032,
       "step": 30,
       "total_flos": 0.0,
-      "train_loss": -0.002354813553392887,
-      "train_runtime": 733.4112,
-      "train_samples_per_second": 1.309,
-      "train_steps_per_second": 0.041
+      "train_loss": 0.0008721697144210338,
+      "train_runtime": 2002.6709,
+      "train_samples_per_second": 0.24,
+      "train_steps_per_second": 0.015
     }
   ],
   "logging_steps": 5,
@@ -134,7 +134,7 @@
     }
   },
   "total_flos": 0.0,
-  "train_batch_size": 4,
+  "train_batch_size": 2,
   "trial_name": null,
   "trial_params": null
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b54993a40d108f108aee06b404faf7282e71ad6feada81ab7f888a425e9051c6
-size 7928
+oid sha256:7ce212c5204dd486d663714791f972520601089fe064d65c6f7d1b24d7d5cb22
+size 6520