Model save

README.md

@@ -1,11 +1,9 @@
 ---
 base_model: Qwen/Qwen2.5-1.5B-Instruct
-datasets: HuggingFaceH4/Bespoke-Stratos-17k
 library_name: transformers
 model_name: Qwen2.5-1.5B-Open-R1-Distill
 tags:
 - generated_from_trainer
-- open-r1
 - trl
 - sft
 licence: license
@@ -13,7 +11,7 @@ licence: license
 
 # Model Card for Qwen2.5-1.5B-Open-R1-Distill
 
-This model is a fine-tuned version of [Qwen/Qwen2.5-1.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct) on the [HuggingFaceH4/Bespoke-Stratos-17k](https://huggingface.co/datasets/HuggingFaceH4/Bespoke-Stratos-17k) dataset.
+This model is a fine-tuned version of [Qwen/Qwen2.5-1.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-1.5B-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/llm-m_wandb-weblab/Qwen2.5-1.5B/runs/b8odb8mc) 
+[<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/llm-m_wandb-weblab/Qwen2.5-1.5B/runs/op7h76qk) 
 
 
 This model was trained with SFT.

all_results.json

@@ -5,10 +5,10 @@
     "eval_samples": 100,
     "eval_samples_per_second": 161.504,
     "eval_steps_per_second": 5.047,
-    "total_flos": 76916824473600.0,
-    "train_loss": 0.8026494443769285,
-    "train_runtime": 500.9768,
+    "total_flos": 76973799899136.0,
+    "train_loss": 0.8025016911636443,
+    "train_runtime": 750.6152,
     "train_samples": 16610,
-    "train_samples_per_second": 43.136,
-    "train_steps_per_second": 0.337
+    "train_samples_per_second": 28.807,
+    "train_steps_per_second": 0.225
 }

config.json

@@ -23,7 +23,7 @@
   "tie_word_embeddings": true,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.49.0.dev0",
-  "use_cache": true,
+  "use_cache": false,
   "use_sliding_window": false,
   "vocab_size": 151936
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1260fda8a12b3724ea50f2d94d94d6b1d67039e55f8ec9ee538d5cac0eaf0072
+oid sha256:1c54337c035c95cdd4fd6ad2ef63dd3251ab63d2f348609987fa76b0a05587b0
 size 3087467144

train_results.json

@@ -1,9 +1,8 @@
 {
-    "epoch": 1.0,
-    "total_flos": 76916824473600.0,
-    "train_loss": 0.8026494443769285,
-    "train_runtime": 500.9768,
+    "total_flos": 76973799899136.0,
+    "train_loss": 0.8025016911636443,
+    "train_runtime": 750.6152,
     "train_samples": 16610,
-    "train_samples_per_second": 43.136,
-    "train_steps_per_second": 0.337
+    "train_samples_per_second": 28.807,
+    "train_steps_per_second": 0.225
 }

trainer_state.json

@@ -10,251 +10,286 @@
   "log_history": [
     {
       "epoch": 0.029585798816568046,
-      "grad_norm": 2.3441456641162017,
+      "grad_norm": 2.297424387110319,
       "learning_rate": 5.882352941176471e-06,
-      "loss": 1.0993,
+      "loss": 1.1002,
+      "mean_token_accuracy": 0.710122095857212,
       "step": 5
     },
     {
       "epoch": 0.05917159763313609,
-      "grad_norm": 1.633349804365292,
+      "grad_norm": 1.5962459582070525,
       "learning_rate": 1.1764705882352942e-05,
-      "loss": 1.0403,
+      "loss": 1.0326,
+      "mean_token_accuracy": 0.7201974025955006,
       "step": 10
     },
     {
       "epoch": 0.08875739644970414,
-      "grad_norm": 0.8842564361880783,
+      "grad_norm": 0.84813226022544,
       "learning_rate": 1.7647058823529414e-05,
-      "loss": 0.9533,
+      "loss": 0.9517,
+      "mean_token_accuracy": 0.7326376870412972,
       "step": 15
     },
     {
       "epoch": 0.11834319526627218,
-      "grad_norm": 0.6055203960588147,
+      "grad_norm": 0.6449426089322932,
       "learning_rate": 1.9980782984658682e-05,
-      "loss": 0.8922,
+      "loss": 0.8804,
+      "mean_token_accuracy": 0.7473364378856827,
       "step": 20
     },
     {
       "epoch": 0.14792899408284024,
-      "grad_norm": 0.5402463110888628,
+      "grad_norm": 0.5511541331468433,
       "learning_rate": 1.9863613034027224e-05,
-      "loss": 0.8552,
+      "loss": 0.8536,
+      "mean_token_accuracy": 0.7518443039297401,
       "step": 25
     },
     {
       "epoch": 0.17751479289940827,
-      "grad_norm": 0.4258701840680344,
+      "grad_norm": 0.442628562609923,
       "learning_rate": 1.9641197940012136e-05,
-      "loss": 0.8283,
+      "loss": 0.8404,
+      "mean_token_accuracy": 0.7545613836853171,
       "step": 30
     },
     {
       "epoch": 0.20710059171597633,
-      "grad_norm": 0.37541501055321486,
+      "grad_norm": 0.3950445105792925,
       "learning_rate": 1.9315910880512792e-05,
-      "loss": 0.8229,
+      "loss": 0.8212,
+      "mean_token_accuracy": 0.7580693317680411,
       "step": 35
     },
     {
       "epoch": 0.23668639053254437,
-      "grad_norm": 0.3981049785839041,
+      "grad_norm": 0.38499101012568593,
       "learning_rate": 1.8891222681391853e-05,
-      "loss": 0.8225,
+      "loss": 0.805,
+      "mean_token_accuracy": 0.7621003697400892,
       "step": 40
     },
     {
       "epoch": 0.26627218934911245,
-      "grad_norm": 0.3442641995819869,
+      "grad_norm": 0.426957308734079,
       "learning_rate": 1.8371664782625287e-05,
-      "loss": 0.8073,
+      "loss": 0.7984,
+      "mean_token_accuracy": 0.763465327169843,
       "step": 45
     },
     {
       "epoch": 0.2958579881656805,
-      "grad_norm": 0.334462527994632,
+      "grad_norm": 0.3931886133887098,
       "learning_rate": 1.7762780887657576e-05,
-      "loss": 0.7977,
+      "loss": 0.7937,
+      "mean_token_accuracy": 0.7642841154247401,
       "step": 50
     },
     {
       "epoch": 0.3254437869822485,
-      "grad_norm": 0.3680761193201223,
+      "grad_norm": 0.37612205413583966,
       "learning_rate": 1.7071067811865477e-05,
-      "loss": 0.7877,
+      "loss": 0.7894,
+      "mean_token_accuracy": 0.7649155785682786,
       "step": 55
     },
     {
       "epoch": 0.35502958579881655,
-      "grad_norm": 0.3692193166132139,
+      "grad_norm": 0.3690088505249562,
       "learning_rate": 1.6303906161279554e-05,
-      "loss": 0.7981,
+      "loss": 0.7999,
+      "mean_token_accuracy": 0.761950455440128,
       "step": 60
     },
     {
       "epoch": 0.38461538461538464,
-      "grad_norm": 0.34448390578574,
+      "grad_norm": 0.3968668346625297,
       "learning_rate": 1.5469481581224274e-05,
-      "loss": 0.7721,
+      "loss": 0.7867,
+      "mean_token_accuracy": 0.764712581061752,
       "step": 65
     },
     {
       "epoch": 0.41420118343195267,
-      "grad_norm": 0.3452861149232402,
+      "grad_norm": 0.37433783603713877,
       "learning_rate": 1.4576697415156818e-05,
-      "loss": 0.7741,
+      "loss": 0.7782,
+      "mean_token_accuracy": 0.7675159747533142,
       "step": 70
     },
     {
       "epoch": 0.4437869822485207,
-      "grad_norm": 0.3363364574541925,
+      "grad_norm": 0.33916682700714146,
       "learning_rate": 1.3635079705638298e-05,
-      "loss": 0.7852,
+      "loss": 0.777,
+      "mean_token_accuracy": 0.7675317758209201,
       "step": 75
     },
     {
       "epoch": 0.47337278106508873,
-      "grad_norm": 0.35158679411054844,
+      "grad_norm": 0.34679361649836743,
       "learning_rate": 1.2654675551080724e-05,
-      "loss": 0.7666,
+      "loss": 0.7783,
+      "mean_token_accuracy": 0.7665991894058249,
       "step": 80
     },
     {
       "epoch": 0.5029585798816568,
-      "grad_norm": 0.3604023099826107,
+      "grad_norm": 0.3364148123003455,
       "learning_rate": 1.164594590280734e-05,
-      "loss": 0.7706,
+      "loss": 0.7785,
+      "mean_token_accuracy": 0.7670094150622615,
       "step": 85
     },
     {
       "epoch": 0.5325443786982249,
-      "grad_norm": 0.35782845320303824,
+      "grad_norm": 0.3495734033672351,
       "learning_rate": 1.0619653946285948e-05,
-      "loss": 0.77,
+      "loss": 0.7723,
+      "mean_token_accuracy": 0.7682055741851116,
       "step": 90
     },
     {
       "epoch": 0.5621301775147929,
-      "grad_norm": 0.382442839254331,
+      "grad_norm": 0.3432574047378308,
       "learning_rate": 9.586750257511868e-06,
-      "loss": 0.7771,
+      "loss": 0.7669,
+      "mean_token_accuracy": 0.7697794763971556,
       "step": 95
     },
     {
       "epoch": 0.591715976331361,
-      "grad_norm": 0.3330090428360514,
+      "grad_norm": 0.34082459115638813,
       "learning_rate": 8.558255959926533e-06,
-      "loss": 0.753,
+      "loss": 0.7582,
+      "mean_token_accuracy": 0.7714136401538525,
       "step": 100
     },
     {
       "epoch": 0.591715976331361,
-      "eval_loss": 0.7880030870437622,
-      "eval_runtime": 0.8026,
-      "eval_samples_per_second": 159.487,
-      "eval_steps_per_second": 4.984,
+      "eval_loss": 0.7885270118713379,
+      "eval_mean_token_accuracy": 0.7488056872787593,
+      "eval_runtime": 2.2705,
+      "eval_samples_per_second": 56.816,
+      "eval_steps_per_second": 2.202,
       "step": 100
     },
     {
       "epoch": 0.621301775147929,
-      "grad_norm": 0.33382381973562353,
+      "grad_norm": 0.3500865265546408,
       "learning_rate": 7.545145128592009e-06,
-      "loss": 0.7569,
+      "loss": 0.759,
+      "mean_token_accuracy": 0.7715606650120836,
       "step": 105
     },
     {
       "epoch": 0.650887573964497,
-      "grad_norm": 0.342407693906868,
+      "grad_norm": 0.3153601426300555,
       "learning_rate": 6.558227696373617e-06,
-      "loss": 0.7681,
+      "loss": 0.7775,
+      "mean_token_accuracy": 0.7664943003135225,
       "step": 110
     },
     {
       "epoch": 0.6804733727810651,
-      "grad_norm": 0.30933337277908973,
+      "grad_norm": 0.3077751865894769,
       "learning_rate": 5.608034111526298e-06,
-      "loss": 0.7623,
+      "loss": 0.7636,
+      "mean_token_accuracy": 0.7702647251741424,
       "step": 115
     },
     {
       "epoch": 0.7100591715976331,
-      "grad_norm": 0.2990413611811013,
+      "grad_norm": 0.2978423880943666,
       "learning_rate": 4.704702977392914e-06,
-      "loss": 0.7514,
+      "loss": 0.7511,
+      "mean_token_accuracy": 0.7742831183944568,
       "step": 120
     },
     {
       "epoch": 0.7396449704142012,
-      "grad_norm": 0.34349997716613806,
+      "grad_norm": 0.3160810225440028,
       "learning_rate": 3.857872873103322e-06,
-      "loss": 0.7538,
+      "loss": 0.7594,
+      "mean_token_accuracy": 0.7718046260565352,
       "step": 125
     },
     {
       "epoch": 0.7692307692307693,
-      "grad_norm": 0.3113170176717584,
+      "grad_norm": 0.3193177717083944,
       "learning_rate": 3.0765795095517026e-06,
-      "loss": 0.7555,
+      "loss": 0.7623,
+      "mean_token_accuracy": 0.7706338288914129,
       "step": 130
     },
     {
       "epoch": 0.7988165680473372,
-      "grad_norm": 0.3230681953500364,
+      "grad_norm": 0.301902197424481,
       "learning_rate": 2.369159318001937e-06,
-      "loss": 0.7584,
+      "loss": 0.7684,
+      "mean_token_accuracy": 0.7688698616474955,
       "step": 135
     },
     {
       "epoch": 0.8284023668639053,
-      "grad_norm": 0.2991556070852993,
+      "grad_norm": 0.2941951800393329,
       "learning_rate": 1.743160500034443e-06,
-      "loss": 0.7498,
+      "loss": 0.7596,
+      "mean_token_accuracy": 0.7716673849461525,
       "step": 140
     },
     {
       "epoch": 0.8579881656804734,
-      "grad_norm": 0.3114696672768526,
+      "grad_norm": 0.3117346710739024,
       "learning_rate": 1.2052624879351105e-06,
-      "loss": 0.7566,
+      "loss": 0.7595,
+      "mean_token_accuracy": 0.7714754205949017,
       "step": 145
     },
     {
       "epoch": 0.8875739644970414,
-      "grad_norm": 0.2949319010398544,
+      "grad_norm": 0.2910628005200283,
       "learning_rate": 7.612046748871327e-07,
-      "loss": 0.7666,
+      "loss": 0.7505,
+      "mean_token_accuracy": 0.7739860912523475,
       "step": 150
     },
     {
       "epoch": 0.9171597633136095,
-      "grad_norm": 0.29707992694247826,
+      "grad_norm": 0.28002315513328135,
       "learning_rate": 4.1572517541747294e-07,
-      "loss": 0.7613,
+      "loss": 0.741,
+      "mean_token_accuracy": 0.7767315836804676,
       "step": 155
     },
     {
       "epoch": 0.9467455621301775,
-      "grad_norm": 0.2858348422013152,
+      "grad_norm": 0.2957619359463069,
       "learning_rate": 1.7251026952640583e-07,
-      "loss": 0.7607,
+      "loss": 0.754,
+      "mean_token_accuracy": 0.7732596586427178,
       "step": 160
     },
     {
       "epoch": 0.9763313609467456,
-      "grad_norm": 0.27606916608468957,
+      "grad_norm": 0.2762985231070698,
       "learning_rate": 3.4155069933301535e-08,
-      "loss": 0.7445,
+      "loss": 0.7467,
+      "mean_token_accuracy": 0.7749716424638167,
       "step": 165
     },
     {
       "epoch": 1.0,
+      "mean_token_accuracy": 0.7704617250967111,
       "step": 169,
-      "total_flos": 76916824473600.0,
-      "train_loss": 0.8026494443769285,
-      "train_runtime": 500.9768,
-      "train_samples_per_second": 43.136,
-      "train_steps_per_second": 0.337
+      "total_flos": 76973799899136.0,
+      "train_loss": 0.8025016911636443,
+      "train_runtime": 750.6152,
+      "train_samples_per_second": 28.807,
+      "train_steps_per_second": 0.225
     }
   ],
   "logging_steps": 5,
@@ -274,8 +309,8 @@
       "attributes": {}
     }
   },
-  "total_flos": 76916824473600.0,
-  "train_batch_size": 4,
+  "total_flos": 76973799899136.0,
+  "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 sha256:44893394a0eba7fd46589a49f8774d43c799edd92b85f27ff4f5ab8717375e95
+oid sha256:4745bf1543ad52003139f9be07590fe262ca4314dff5d0f7b2e7d7f042b61faa
 size 7352