Model save

README.md

@@ -1,19 +1,17 @@
 ---
 base_model: Qwen/Qwen2.5-0.5B-Instruct
-datasets: YangZhoumill/bestofn
 library_name: transformers
-model_name: Qwen2.5-0.5B-Instruct-4230041
+model_name: Qwen2.5-0.5B-Instruct
 tags:
 - generated_from_trainer
-- open-r1
 - trl
 - sft
 licence: license
 ---
 
-# Model Card for Qwen2.5-0.5B-Instruct-4230041
+# Model Card for Qwen2.5-0.5B-Instruct
 
-This model is a fine-tuned version of [Qwen/Qwen2.5-0.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-0.5B-Instruct) on the [YangZhoumill/bestofn](https://huggingface.co/datasets/YangZhoumill/bestofn) dataset.
+This model is a fine-tuned version of [Qwen/Qwen2.5-0.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-0.5B-Instruct).
 It has been trained using [TRL](https://github.com/huggingface/trl).
 
 ## Quick start
@@ -29,7 +27,7 @@ print(output["generated_text"])
 
 ## Training procedure
 
-[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/stevenzhou0816100/huggingface/runs/vv5r4n2p) 
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/stevenzhou0816100/huggingface/runs/hq1anqyw) 
 
 
 This model was trained with SFT.

all_results.json

@@ -1,8 +1,8 @@
 {
-    "total_flos": 5.27742227447808e+16,
-    "train_loss": 0.6343651420258462,
-    "train_runtime": 365.6281,
-    "train_samples": 11477,
-    "train_samples_per_second": 4.103,
-    "train_steps_per_second": 0.514
+    "total_flos": 3.486616950459597e+16,
+    "train_loss": 0.9192673854827881,
+    "train_runtime": 250.2316,
+    "train_samples": 7473,
+    "train_samples_per_second": 3.996,
+    "train_steps_per_second": 0.5
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "total_flos": 5.27742227447808e+16,
-    "train_loss": 0.6343651420258462,
-    "train_runtime": 365.6281,
-    "train_samples": 11477,
-    "train_samples_per_second": 4.103,
-    "train_steps_per_second": 0.514
+    "total_flos": 3.486616950459597e+16,
+    "train_loss": 0.9192673854827881,
+    "train_runtime": 250.2316,
+    "train_samples": 7473,
+    "train_samples_per_second": 3.996,
+    "train_steps_per_second": 0.5
 }

trainer_state.json

@@ -4,282 +4,198 @@
   "best_model_checkpoint": null,
   "epoch": 1.0,
   "eval_steps": 500,
-  "global_step": 188,
+  "global_step": 125,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.026595744680851064,
-      "grad_norm": 67.0,
-      "learning_rate": 2e-05,
-      "loss": 2.9921,
+      "epoch": 0.04,
+      "grad_norm": 62.0,
+      "learning_rate": 2.857142857142857e-05,
+      "loss": 3.2566,
       "step": 5
     },
     {
-      "epoch": 0.05319148936170213,
-      "grad_norm": 11.1875,
-      "learning_rate": 4.5e-05,
-      "loss": 1.2994,
+      "epoch": 0.08,
+      "grad_norm": 9.625,
+      "learning_rate": 4.996811065272715e-05,
+      "loss": 1.5876,
       "step": 10
     },
     {
-      "epoch": 0.0797872340425532,
-      "grad_norm": 7.75,
-      "learning_rate": 4.9943953065892183e-05,
-      "loss": 0.9284,
+      "epoch": 0.12,
+      "grad_norm": 7.28125,
+      "learning_rate": 4.9610392803331726e-05,
+      "loss": 1.2605,
       "step": 15
     },
     {
-      "epoch": 0.10638297872340426,
-      "grad_norm": 7.125,
-      "learning_rate": 4.97167408627609e-05,
-      "loss": 0.9604,
+      "epoch": 0.16,
+      "grad_norm": 49.75,
+      "learning_rate": 4.8861446190538576e-05,
+      "loss": 1.1706,
       "step": 20
     },
     {
-      "epoch": 0.13297872340425532,
-      "grad_norm": 5.40625,
-      "learning_rate": 4.931662736686749e-05,
-      "loss": 0.8497,
+      "epoch": 0.2,
+      "grad_norm": 8.8125,
+      "learning_rate": 4.7734522928852436e-05,
+      "loss": 1.0933,
       "step": 25
     },
     {
-      "epoch": 0.1595744680851064,
-      "grad_norm": 8.3125,
-      "learning_rate": 4.8746726452615006e-05,
-      "loss": 0.835,
+      "epoch": 0.24,
+      "grad_norm": 9.75,
+      "learning_rate": 4.624956317935659e-05,
+      "loss": 1.0288,
       "step": 30
     },
     {
-      "epoch": 0.18617021276595744,
-      "grad_norm": 11.1875,
-      "learning_rate": 4.801147336122219e-05,
-      "loss": 0.7549,
+      "epoch": 0.28,
+      "grad_norm": 7.59375,
+      "learning_rate": 4.443284232176311e-05,
+      "loss": 0.9685,
       "step": 35
     },
     {
-      "epoch": 0.2127659574468085,
-      "grad_norm": 7.78125,
-      "learning_rate": 4.7116590183557117e-05,
-      "loss": 0.7351,
+      "epoch": 0.32,
+      "grad_norm": 10.4375,
+      "learning_rate": 4.2316506028963374e-05,
+      "loss": 0.89,
       "step": 40
     },
     {
-      "epoch": 0.2393617021276596,
-      "grad_norm": 9.375,
-      "learning_rate": 4.6069041328092e-05,
-      "loss": 0.6336,
+      "epoch": 0.36,
+      "grad_norm": 9.625,
+      "learning_rate": 3.993800147062685e-05,
+      "loss": 0.8948,
       "step": 45
     },
     {
-      "epoch": 0.26595744680851063,
-      "grad_norm": 10.625,
-      "learning_rate": 4.487697932054875e-05,
-      "loss": 0.7065,
+      "epoch": 0.4,
+      "grad_norm": 16.25,
+      "learning_rate": 3.733941471032425e-05,
+      "loss": 0.85,
       "step": 50
     },
     {
-      "epoch": 0.2925531914893617,
-      "grad_norm": 394.0,
-      "learning_rate": 4.354968135704856e-05,
-      "loss": 0.7266,
+      "epoch": 0.44,
+      "grad_norm": 59.25,
+      "learning_rate": 3.4566726020493854e-05,
+      "loss": 0.8486,
       "step": 55
     },
     {
-      "epoch": 0.3191489361702128,
-      "grad_norm": 632.0,
-      "learning_rate": 4.209747710453884e-05,
-      "loss": 0.7077,
+      "epoch": 0.48,
+      "grad_norm": 78.5,
+      "learning_rate": 3.1668996291960073e-05,
+      "loss": 0.7863,
       "step": 60
     },
     {
-      "epoch": 0.34574468085106386,
-      "grad_norm": 17.625,
-      "learning_rate": 4.05316683103895e-05,
-      "loss": 0.721,
+      "epoch": 0.52,
+      "grad_norm": 14.5625,
+      "learning_rate": 2.869749893394902e-05,
+      "loss": 0.7641,
       "step": 65
     },
     {
-      "epoch": 0.3723404255319149,
-      "grad_norm": 13.9375,
-      "learning_rate": 3.886444084679482e-05,
-      "loss": 0.5756,
+      "epoch": 0.56,
+      "grad_norm": 9.5,
+      "learning_rate": 2.570481262505563e-05,
+      "loss": 0.7131,
       "step": 70
     },
     {
-      "epoch": 0.39893617021276595,
-      "grad_norm": 14.4375,
-      "learning_rate": 3.710876987449346e-05,
-      "loss": 0.5324,
+      "epoch": 0.6,
+      "grad_norm": 11.25,
+      "learning_rate": 2.2743890968333453e-05,
+      "loss": 0.6848,
       "step": 75
     },
     {
-      "epoch": 0.425531914893617,
-      "grad_norm": 24.0,
-      "learning_rate": 3.527831886386802e-05,
-      "loss": 0.5874,
+      "epoch": 0.64,
+      "grad_norm": 13.0625,
+      "learning_rate": 1.986712551234432e-05,
+      "loss": 0.6762,
       "step": 80
     },
     {
-      "epoch": 0.4521276595744681,
-      "grad_norm": 14.25,
-      "learning_rate": 3.338733325928991e-05,
-      "loss": 0.5228,
+      "epoch": 0.68,
+      "grad_norm": 10.625,
+      "learning_rate": 1.7125418717390167e-05,
+      "loss": 0.643,
       "step": 85
     },
     {
-      "epoch": 0.4787234042553192,
-      "grad_norm": 24.875,
-      "learning_rate": 3.1450529614264715e-05,
-      "loss": 0.5005,
+      "epoch": 0.72,
+      "grad_norm": 13.375,
+      "learning_rate": 1.4567283270175847e-05,
+      "loss": 0.6211,
       "step": 90
     },
     {
-      "epoch": 0.5053191489361702,
-      "grad_norm": 15.75,
-      "learning_rate": 2.9482981060181047e-05,
-      "loss": 0.4858,
+      "epoch": 0.76,
+      "grad_norm": 15.375,
+      "learning_rate": 1.2237983683933638e-05,
+      "loss": 0.6132,
       "step": 95
     },
     {
-      "epoch": 0.5319148936170213,
-      "grad_norm": 21.125,
-      "learning_rate": 2.7500000000000004e-05,
-      "loss": 0.5236,
+      "epoch": 0.8,
+      "grad_norm": 11.9375,
+      "learning_rate": 1.0178735372827107e-05,
+      "loss": 0.6187,
       "step": 100
     },
     {
-      "epoch": 0.5585106382978723,
-      "grad_norm": 10.3125,
-      "learning_rate": 2.551701893981897e-05,
-      "loss": 0.456,
+      "epoch": 0.84,
+      "grad_norm": 23.75,
+      "learning_rate": 8.425975372482405e-06,
+      "loss": 0.6055,
       "step": 105
     },
     {
-      "epoch": 0.5851063829787234,
-      "grad_norm": 15.0,
-      "learning_rate": 2.354947038573529e-05,
-      "loss": 0.4441,
+      "epoch": 0.88,
+      "grad_norm": 13.1875,
+      "learning_rate": 7.010717610764453e-06,
+      "loss": 0.6133,
       "step": 110
     },
     {
-      "epoch": 0.6117021276595744,
-      "grad_norm": 11.1875,
-      "learning_rate": 2.1612666740710093e-05,
-      "loss": 0.4312,
+      "epoch": 0.92,
+      "grad_norm": 12.125,
+      "learning_rate": 5.9580041368548775e-06,
+      "loss": 0.6015,
       "step": 115
     },
     {
-      "epoch": 0.6382978723404256,
-      "grad_norm": 61.5,
-      "learning_rate": 1.972168113613198e-05,
-      "loss": 0.4592,
+      "epoch": 0.96,
+      "grad_norm": 10.1875,
+      "learning_rate": 5.286462018769748e-06,
+      "loss": 0.5994,
       "step": 120
     },
     {
-      "epoch": 0.6648936170212766,
-      "grad_norm": 18.25,
-      "learning_rate": 1.7891230125506537e-05,
-      "loss": 0.4386,
+      "epoch": 1.0,
+      "grad_norm": 12.9375,
+      "learning_rate": 5.007973749722316e-06,
+      "loss": 0.5922,
       "step": 125
     },
-    {
-      "epoch": 0.6914893617021277,
-      "grad_norm": 64.0,
-      "learning_rate": 1.6135559153205187e-05,
-      "loss": 0.4423,
-      "step": 130
-    },
-    {
-      "epoch": 0.7180851063829787,
-      "grad_norm": 10.0,
-      "learning_rate": 1.4468331689610495e-05,
-      "loss": 0.4148,
-      "step": 135
-    },
-    {
-      "epoch": 0.7446808510638298,
-      "grad_norm": 8.1875,
-      "learning_rate": 1.2902522895461167e-05,
-      "loss": 0.3673,
-      "step": 140
-    },
-    {
-      "epoch": 0.7712765957446809,
-      "grad_norm": 7.65625,
-      "learning_rate": 1.1450318642951452e-05,
-      "loss": 0.3963,
-      "step": 145
-    },
-    {
-      "epoch": 0.7978723404255319,
-      "grad_norm": 7.25,
-      "learning_rate": 1.0123020679451254e-05,
-      "loss": 0.4026,
-      "step": 150
-    },
-    {
-      "epoch": 0.824468085106383,
-      "grad_norm": 18.375,
-      "learning_rate": 8.930958671908013e-06,
-      "loss": 0.4144,
-      "step": 155
-    },
-    {
-      "epoch": 0.851063829787234,
-      "grad_norm": 8.5,
-      "learning_rate": 7.883409816442882e-06,
-      "loss": 0.3811,
-      "step": 160
-    },
-    {
-      "epoch": 0.8776595744680851,
-      "grad_norm": 14.3125,
-      "learning_rate": 6.988526638777812e-06,
-      "loss": 0.4083,
-      "step": 165
-    },
-    {
-      "epoch": 0.9042553191489362,
-      "grad_norm": 34.75,
-      "learning_rate": 6.253273547384997e-06,
-      "loss": 0.4174,
-      "step": 170
-    },
-    {
-      "epoch": 0.9308510638297872,
-      "grad_norm": 13.3125,
-      "learning_rate": 5.683372633132516e-06,
-      "loss": 0.3967,
-      "step": 175
-    },
-    {
-      "epoch": 0.9574468085106383,
-      "grad_norm": 14.25,
-      "learning_rate": 5.283259137239111e-06,
-      "loss": 0.3973,
-      "step": 180
-    },
-    {
-      "epoch": 0.9840425531914894,
-      "grad_norm": 18.125,
-      "learning_rate": 5.056046934107815e-06,
-      "loss": 0.3972,
-      "step": 185
-    },
     {
       "epoch": 1.0,
-      "step": 188,
-      "total_flos": 5.27742227447808e+16,
-      "train_loss": 0.6343651420258462,
-      "train_runtime": 365.6281,
-      "train_samples_per_second": 4.103,
-      "train_steps_per_second": 0.514
+      "step": 125,
+      "total_flos": 3.486616950459597e+16,
+      "train_loss": 0.9192673854827881,
+      "train_runtime": 250.2316,
+      "train_samples_per_second": 3.996,
+      "train_steps_per_second": 0.5
     }
   ],
   "logging_steps": 5,
-  "max_steps": 188,
+  "max_steps": 125,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 1,
   "save_steps": 500,
@@ -295,7 +211,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 5.27742227447808e+16,
+  "total_flos": 3.486616950459597e+16,
   "train_batch_size": 2,
   "trial_name": null,
   "trial_params": null