Model save

README.md

@@ -1,11 +1,9 @@
 ---
 base_model: KAKA22/CodeRM-8B
-datasets: zzzzit/taco3
 library_name: transformers
 model_name: coderm-8b
 tags:
 - generated_from_trainer
-- open-r1
 - trl
 - grpo
 licence: license
@@ -13,7 +11,7 @@ licence: license
 
 # Model Card for coderm-8b
 
-This model is a fine-tuned version of [KAKA22/CodeRM-8B](https://huggingface.co/KAKA22/CodeRM-8B) on the [zzzzit/taco3](https://huggingface.co/datasets/zzzzit/taco3) dataset.
+This model is a fine-tuned version of [KAKA22/CodeRM-8B](https://huggingface.co/KAKA22/CodeRM-8B).
 It has been trained using [TRL](https://github.com/huggingface/trl).
 
 ## Quick start

adapter_config.json

@@ -29,13 +29,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "up_proj",
-    "q_proj",
-    "k_proj",
     "gate_proj",
+    "q_proj",
     "o_proj",
-    "down_proj",
-    "v_proj"
+    "v_proj",
+    "k_proj",
+    "up_proj",
+    "down_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b1cff78d208a0464979b427750cf2dccacb49a08960467848e37ab56fee24a15
+oid sha256:f065bcd1c2467f3bbc7b68bf2a2b5a2b8f9f44eb9e9cf72845c4265b8dbed92a
 size 83946192

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 0.0021783057984875913,
-    "train_runtime": 33449.6769,
+    "train_loss": 0.0,
+    "train_runtime": 0.1209,
     "train_samples": 1315,
-    "train_samples_per_second": 0.079,
-    "train_steps_per_second": 0.007
+    "train_samples_per_second": 10872.756,
+    "train_steps_per_second": 909.508
 }

tokenizer.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:65ff5472d095ccd9332d9e723153d7bc7226cb6be9c1bffda738b5ba2e71bf26
-size 17210084
+oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b
+size 17209920

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 0.0021783057984875913,
-    "train_runtime": 33449.6769,
+    "train_loss": 0.0,
+    "train_runtime": 0.1209,
     "train_samples": 1315,
-    "train_samples_per_second": 0.079,
-    "train_steps_per_second": 0.007
+    "train_samples_per_second": 10872.756,
+    "train_steps_per_second": 909.508
 }

trainer_state.json

@@ -2,9 +2,9 @@
   "best_global_step": null,
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.994582264043342,
+  "epoch": 1.8303393213572854,
   "eval_steps": 500,
-  "global_step": 218,
+  "global_step": 200,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -5810,541 +5810,19 @@
       "step": 200
     },
     {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -4.984375,
-      "completions/max_length": 3584.0,
-      "completions/max_terminated_length": 1500.0,
-      "completions/mean_length": 541.9375,
-      "completions/mean_terminated_length": 533.9948120117188,
-      "completions/min_length": 277.0,
-      "completions/min_terminated_length": 277.0,
-      "entropy": 0.1104888916015625,
-      "epoch": 1.8394639292842885,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.03310083597898483,
-      "kl": 0.004993915557861328,
-      "learning_rate": 4.992888225905467e-07,
-      "loss": 0.0189,
-      "num_tokens": 82488820.0,
-      "reward": 0.9836899042129517,
-      "reward_std": 0.3145735263824463,
-      "rewards/cross_solution_unittest_reward_v2/mean": 0.9836898446083069,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.7407367825508118,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 201
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -5.0,
-      "completions/max_length": 1000.0,
-      "completions/max_terminated_length": 1000.0,
-      "completions/mean_length": 516.8333740234375,
-      "completions/mean_terminated_length": 516.8333740234375,
-      "completions/min_length": 277.0,
-      "completions/min_terminated_length": 277.0,
-      "entropy": 0.1177825927734375,
-      "epoch": 1.8485885372112918,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.03865882754325867,
-      "kl": 0.005397796630859375,
-      "learning_rate": 4.509775855592613e-07,
-      "loss": 0.0014,
-      "num_tokens": 82897596.0,
-      "reward": 1.1494131088256836,
-      "reward_std": 0.3502509593963623,
-      "rewards/cross_solution_unittest_reward_v2/mean": 1.149412989616394,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.5175426006317139,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 202
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -5.0,
-      "completions/max_length": 1046.0,
-      "completions/max_terminated_length": 1046.0,
-      "completions/mean_length": 489.6536560058594,
-      "completions/mean_terminated_length": 489.6536560058594,
-      "completions/min_length": 208.0,
-      "completions/min_terminated_length": 208.0,
-      "entropy": 0.125946044921875,
-      "epoch": 1.8577131451382949,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.040030594915151596,
-      "kl": 0.005590915679931641,
-      "learning_rate": 4.0507026385502747e-07,
-      "loss": -0.0072,
-      "num_tokens": 83270287.0,
-      "reward": 1.2290350198745728,
-      "reward_std": 0.35159873962402344,
-      "rewards/cross_solution_unittest_reward_v2/mean": 1.2290349006652832,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.4783675968647003,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 203
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -4.875,
-      "completions/max_length": 3584.0,
-      "completions/max_terminated_length": 3505.0,
-      "completions/mean_length": 660.6927490234375,
-      "completions/mean_terminated_length": 598.4946899414062,
-      "completions/min_length": 235.0,
-      "completions/min_terminated_length": 235.0,
-      "entropy": 0.123809814453125,
-      "epoch": 1.866837753065298,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.0407257154583931,
-      "kl": 0.00537109375,
-      "learning_rate": 3.615784144005796e-07,
-      "loss": 0.0309,
-      "num_tokens": 83734721.0,
-      "reward": 1.1413307189941406,
-      "reward_std": 0.337158203125,
-      "rewards/cross_solution_unittest_reward_v2/mean": 1.1413307189941406,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.5994848012924194,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 204
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -5.0,
-      "completions/max_length": 1415.0,
-      "completions/max_terminated_length": 1415.0,
-      "completions/mean_length": 505.7630310058594,
-      "completions/mean_terminated_length": 505.7630310058594,
-      "completions/min_length": 195.0,
-      "completions/min_terminated_length": 195.0,
-      "entropy": 0.1149749755859375,
-      "epoch": 1.875962360992301,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.041536130011081696,
-      "kl": 0.005519866943359375,
-      "learning_rate": 3.2051298603643754e-07,
-      "loss": 0.0015,
-      "num_tokens": 84184366.0,
-      "reward": 1.01891028881073,
-      "reward_std": 0.2728710174560547,
-      "rewards/cross_solution_unittest_reward_v2/mean": 1.0189101696014404,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.6233912110328674,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 205
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -4.984375,
-      "completions/max_length": 3584.0,
-      "completions/max_terminated_length": 924.0,
-      "completions/mean_length": 473.2578125,
-      "completions/mean_terminated_length": 465.1357727050781,
-      "completions/min_length": 249.0,
-      "completions/min_terminated_length": 249.0,
-      "entropy": 0.1327056884765625,
-      "epoch": 1.8850869689193042,
-      "frac_reward_zero_std": 0.0833333358168602,
-      "grad_norm": 0.03881911188364029,
-      "kl": 0.005913257598876953,
-      "learning_rate": 2.818843167645835e-07,
-      "loss": 0.0182,
-      "num_tokens": 84562041.0,
-      "reward": 1.165932297706604,
-      "reward_std": 0.26059168577194214,
-      "rewards/cross_solution_unittest_reward_v2/mean": 1.1659321784973145,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.5712704062461853,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 206
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -5.0,
-      "completions/max_length": 1047.0,
-      "completions/max_terminated_length": 1047.0,
-      "completions/mean_length": 495.7161560058594,
-      "completions/mean_terminated_length": 495.7161560058594,
-      "completions/min_length": 240.0,
-      "completions/min_terminated_length": 240.0,
-      "entropy": 0.1186676025390625,
-      "epoch": 1.8942115768463075,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.04146021604537964,
-      "kl": 0.006419658660888672,
-      "learning_rate": 2.4570213114592957e-07,
-      "loss": 0.0104,
-      "num_tokens": 84935028.0,
-      "reward": 0.9594624042510986,
-      "reward_std": 0.42620453238487244,
-      "rewards/cross_solution_unittest_reward_v2/mean": 0.9594623446464539,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.6278195381164551,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 207
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -5.0,
-      "completions/max_length": 1099.0,
-      "completions/max_terminated_length": 1099.0,
-      "completions/mean_length": 495.6770935058594,
-      "completions/mean_terminated_length": 495.6770935058594,
-      "completions/min_length": 255.0,
-      "completions/min_terminated_length": 255.0,
-      "entropy": 0.1346588134765625,
-      "epoch": 1.9033361847733106,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.04779229685664177,
-      "kl": 0.008728981018066406,
-      "learning_rate": 2.119755378522137e-07,
-      "loss": -0.0073,
-      "num_tokens": 85355856.0,
-      "reward": 0.9647444486618042,
-      "reward_std": 0.30025237798690796,
-      "rewards/cross_solution_unittest_reward_v2/mean": 0.9647443890571594,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.5932243466377258,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 208
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -5.0,
-      "completions/max_length": 1214.0,
-      "completions/max_terminated_length": 1214.0,
-      "completions/mean_length": 540.359375,
-      "completions/mean_terminated_length": 540.359375,
-      "completions/min_length": 268.0,
-      "completions/min_terminated_length": 268.0,
-      "entropy": 0.1104736328125,
-      "epoch": 1.9124607927003137,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.037702981382608414,
-      "kl": 0.005553245544433594,
-      "learning_rate": 1.8071302737293294e-07,
-      "loss": 0.0006,
-      "num_tokens": 85814954.0,
-      "reward": 0.8188127875328064,
-      "reward_std": 0.3269250988960266,
-      "rewards/cross_solution_unittest_reward_v2/mean": 0.8188127875328064,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.49607712030410767,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 209
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -4.984375,
-      "completions/max_length": 3584.0,
-      "completions/max_terminated_length": 878.0,
-      "completions/mean_length": 481.09637451171875,
-      "completions/mean_terminated_length": 472.9947814941406,
-      "completions/min_length": 239.0,
-      "completions/min_terminated_length": 239.0,
-      "entropy": 0.1308746337890625,
-      "epoch": 1.9215854006273168,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.04410184547305107,
-      "kl": 0.0059986114501953125,
-      "learning_rate": 1.519224698779198e-07,
-      "loss": 0.002,
-      "num_tokens": 86181407.0,
-      "reward": 1.0832650661468506,
-      "reward_std": 0.34354862570762634,
-      "rewards/cross_solution_unittest_reward_v2/mean": 1.083264946937561,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.59918212890625,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 210
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -4.984375,
-      "completions/max_length": 3584.0,
-      "completions/max_terminated_length": 2692.0,
-      "completions/mean_length": 491.2083435058594,
-      "completions/mean_terminated_length": 483.1331787109375,
-      "completions/min_length": 259.0,
-      "completions/min_terminated_length": 259.0,
-      "entropy": 0.132293701171875,
-      "epoch": 1.9307100085543198,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.0377914234995842,
-      "kl": 0.007014751434326172,
-      "learning_rate": 1.2561111323605714e-07,
-      "loss": 0.0029,
-      "num_tokens": 86610895.0,
-      "reward": 1.3885529041290283,
-      "reward_std": 0.29931795597076416,
-      "rewards/cross_solution_unittest_reward_v2/mean": 1.3885527849197388,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.48529261350631714,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 211
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -4.9375,
-      "completions/max_length": 3584.0,
-      "completions/max_terminated_length": 2171.0,
-      "completions/mean_length": 559.4271240234375,
-      "completions/mean_terminated_length": 527.5894775390625,
-      "completions/min_length": 262.0,
-      "completions/min_terminated_length": 262.0,
-      "entropy": 0.1161041259765625,
-      "epoch": 1.9398346164813232,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.03503241017460823,
-      "kl": 0.004271984100341797,
-      "learning_rate": 1.0178558119067316e-07,
-      "loss": 0.0119,
-      "num_tokens": 87075387.0,
-      "reward": 1.2032862901687622,
-      "reward_std": 0.2668260335922241,
-      "rewards/cross_solution_unittest_reward_v2/mean": 1.2032862901687622,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.6290963888168335,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 212
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -5.0,
-      "completions/max_length": 945.0,
-      "completions/max_terminated_length": 945.0,
-      "completions/mean_length": 489.3489685058594,
-      "completions/mean_terminated_length": 489.3489685058594,
-      "completions/min_length": 201.0,
-      "completions/min_terminated_length": 201.0,
-      "entropy": 0.1250457763671875,
-      "epoch": 1.9489592244083263,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.044615913182497025,
-      "kl": 0.006129264831542969,
-      "learning_rate": 8.04518716920466e-08,
-      "loss": 0.0048,
-      "num_tokens": 87472705.0,
-      "reward": 1.369114637374878,
-      "reward_std": 0.3354648947715759,
-      "rewards/cross_solution_unittest_reward_v2/mean": 1.3691145181655884,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.6147169470787048,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 213
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -5.0,
-      "completions/max_length": 968.0,
-      "completions/max_terminated_length": 968.0,
-      "completions/mean_length": 509.12762451171875,
-      "completions/mean_terminated_length": 509.12762451171875,
-      "completions/min_length": 265.0,
-      "completions/min_terminated_length": 265.0,
-      "entropy": 0.11688232421875,
-      "epoch": 1.9580838323353293,
-      "frac_reward_zero_std": 0.1666666716337204,
-      "grad_norm": 0.03762752562761307,
-      "kl": 0.005598545074462891,
-      "learning_rate": 6.161535538745877e-08,
-      "loss": -0.0161,
-      "num_tokens": 87886050.0,
-      "reward": 1.2271226644515991,
-      "reward_std": 0.2903611361980438,
-      "rewards/cross_solution_unittest_reward_v2/mean": 1.2271226644515991,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.6599547863006592,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 214
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -4.984375,
-      "completions/max_length": 3584.0,
-      "completions/max_terminated_length": 1367.0,
-      "completions/mean_length": 520.9193115234375,
-      "completions/mean_terminated_length": 512.9216918945312,
-      "completions/min_length": 221.0,
-      "completions/min_terminated_length": 221.0,
-      "entropy": 0.1234893798828125,
-      "epoch": 1.9672084402623324,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.04658033326268196,
-      "kl": 0.005661964416503906,
-      "learning_rate": 4.528077426915412e-08,
-      "loss": -0.0018,
-      "num_tokens": 88293243.0,
-      "reward": 1.1864874362945557,
-      "reward_std": 0.25069719552993774,
-      "rewards/cross_solution_unittest_reward_v2/mean": 1.1864873170852661,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.6851222515106201,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 215
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -4.921875,
-      "completions/max_length": 3584.0,
-      "completions/max_terminated_length": 3486.0,
-      "completions/mean_length": 562.3151245117188,
-      "completions/mean_terminated_length": 522.4512329101562,
-      "completions/min_length": 265.0,
-      "completions/min_terminated_length": 265.0,
-      "entropy": 0.122833251953125,
-      "epoch": 1.9763330481893355,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.04474826529622078,
-      "kl": 0.005382061004638672,
-      "learning_rate": 3.1452240480577265e-08,
-      "loss": 0.0239,
-      "num_tokens": 88760692.0,
-      "reward": 1.1309950351715088,
-      "reward_std": 0.40037307143211365,
-      "rewards/cross_solution_unittest_reward_v2/mean": 1.1309949159622192,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.6654509902000427,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 216
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -4.984375,
-      "completions/max_length": 3584.0,
-      "completions/max_terminated_length": 1410.0,
-      "completions/mean_length": 521.3411865234375,
-      "completions/mean_terminated_length": 513.3446655273438,
-      "completions/min_length": 243.0,
-      "completions/min_terminated_length": 243.0,
-      "entropy": 0.127471923828125,
-      "epoch": 1.9854576561163388,
-      "frac_reward_zero_std": 0.0833333358168602,
-      "grad_norm": 0.044650688767433167,
-      "kl": 0.006046295166015625,
-      "learning_rate": 2.013323528115674e-08,
-      "loss": -0.002,
-      "num_tokens": 89202935.0,
-      "reward": 0.8677881956100464,
-      "reward_std": 0.3405400514602661,
-      "rewards/cross_solution_unittest_reward_v2/mean": 0.8677881360054016,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.6135694980621338,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 217
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": -4.984375,
-      "completions/max_length": 3584.0,
-      "completions/max_terminated_length": 1028.0,
-      "completions/mean_length": 511.7395935058594,
-      "completions/mean_terminated_length": 503.718017578125,
-      "completions/min_length": 224.0,
-      "completions/min_terminated_length": 224.0,
-      "entropy": 0.125701904296875,
-      "epoch": 1.994582264043342,
-      "frac_reward_zero_std": 0.0833333358168602,
-      "grad_norm": 0.03669625520706177,
-      "kl": 0.0051422119140625,
-      "learning_rate": 1.1326608169920373e-08,
-      "loss": -0.0078,
-      "num_tokens": 89621715.0,
-      "reward": 0.8944796323776245,
-      "reward_std": 0.3476303219795227,
-      "rewards/cross_solution_unittest_reward_v2/mean": 0.8944795727729797,
-      "rewards/cross_solution_unittest_reward_v2/std": 0.6513932943344116,
-      "rewards/format_reward/mean": 0.0,
-      "rewards/format_reward/std": 0.0,
-      "step": 218
-    },
-    {
-      "epoch": 1.994582264043342,
-      "step": 218,
+      "epoch": 1.8303393213572854,
+      "step": 200,
       "total_flos": 0.0,
-      "train_loss": 0.0021783057984875913,
-      "train_runtime": 33449.6769,
-      "train_samples_per_second": 0.079,
-      "train_steps_per_second": 0.007
+      "train_loss": 0.0,
+      "train_runtime": 0.1209,
+      "train_samples_per_second": 10872.756,
+      "train_steps_per_second": 909.508
     }
   ],
   "logging_steps": 1,
-  "max_steps": 220,
-  "num_input_tokens_seen": 89621715,
-  "num_train_epochs": 2,
+  "max_steps": 110,
+  "num_input_tokens_seen": 82038884,
+  "num_train_epochs": 1,
   "save_steps": 25,
   "stateful_callbacks": {
     "TrainerControl": {

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:022b3b919c6ddabd570d1dd58f590cdb1fe6a26233bf58a1069ac95561d820b2
+oid sha256:a0a259caaabc47b967be65bcdc75863a912c37b57ab516debc0baf9684f92f3e
 size 8376