End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:711e814ee1b858386d13dd5432aae25763a9866558d8182f12a53f893015b22b
+oid sha256:7b3f5f4bbadb22efcb0c00e7b95e34d767dca28be587ddc67a3950d9453883cf
 size 35237104

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 5.182587085528212e-06,
-    "train_runtime": 867.121,
-    "train_samples": 20,
-    "train_samples_per_second": 0.369,
-    "train_steps_per_second": 0.023
+    "train_loss": 2.5685094556138212e-06,
+    "train_runtime": 536.6447,
+    "train_samples": 14,
+    "train_samples_per_second": 0.596,
+    "train_steps_per_second": 0.037
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 5.182587085528212e-06,
-    "train_runtime": 867.121,
-    "train_samples": 20,
-    "train_samples_per_second": 0.369,
-    "train_steps_per_second": 0.023
+    "train_loss": 2.5685094556138212e-06,
+    "train_runtime": 536.6447,
+    "train_samples": 14,
+    "train_samples_per_second": 0.596,
+    "train_steps_per_second": 0.037
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 4.0,
+  "epoch": 5.0,
   "eval_steps": 500,
   "global_step": 20,
   "is_hyper_param_search": false,
@@ -9,209 +9,209 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 354.15625,
-      "epoch": 0.4,
-      "grad_norm": 1.9123140573501587,
-      "kl": 0.0015067440326674841,
+      "completion_length": 439.21875,
+      "epoch": 0.5714285714285714,
+      "grad_norm": 5.828135013580322,
+      "kl": 0.0014942500565666705,
       "learning_rate": 4.965903258506806e-07,
       "loss": 0.0,
-      "reward": 1.2610771358013153,
-      "reward_std": 1.0868522808887064,
-      "rewards/concensus_correctness_reward_func": 0.12075000256299973,
-      "rewards/consensus_reward_func": 0.0625,
+      "reward": 0.10679153446108103,
+      "reward_std": 0.6297805532813072,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.5,
-      "rewards/question_recreation_reward_func": 0.3578896429389715,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.20822900370694697,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.21993749681860209,
+      "rewards/xmlcount_reward_func": -0.10143748670816422,
       "step": 2
     },
     {
-      "completion_length": 341.4375,
-      "epoch": 0.8,
-      "grad_norm": 1.9063632488250732,
-      "kl": 0.003322049531561788,
+      "completion_length": 345.7083333333333,
+      "epoch": 1.0,
+      "grad_norm": 23.521379470825195,
+      "kl": 0.003981841291533783,
       "learning_rate": 4.698684378016222e-07,
       "loss": 0.0,
-      "reward": 1.7711269692517817,
-      "reward_std": 1.5434346104739234,
-      "rewards/concensus_correctness_reward_func": 0.4007500149309635,
-      "rewards/consensus_reward_func": 0.25,
+      "reward": 1.658778190612793,
+      "reward_std": 2.896653901785612,
+      "rewards/concensus_correctness_reward_func": 0.8787500013907751,
+      "rewards/consensus_reward_func": 0.08333333333333333,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.5625,
-      "rewards/question_recreation_reward_func": 0.41365833627060056,
+      "rewards/final_correctness_reward_func": 0.4166666666666667,
+      "rewards/question_recreation_reward_func": 0.39598656445741653,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.14421874936670065,
+      "rewards/xmlcount_reward_func": -0.1159583330154419,
       "step": 4
     },
     {
-      "completion_length": 345.90625,
-      "epoch": 1.2,
-      "grad_norm": 12.31548023223877,
-      "kl": 0.013164544332539663,
+      "completion_length": 438.5,
+      "epoch": 1.5714285714285714,
+      "grad_norm": 1.3683266639709473,
+      "kl": 0.0017546406452311203,
       "learning_rate": 4.193203929064353e-07,
       "loss": 0.0,
-      "reward": 1.8812013175338507,
-      "reward_std": 1.6580295152962208,
-      "rewards/concensus_correctness_reward_func": 0.24287500232458115,
-      "rewards/consensus_reward_func": 0.125,
+      "reward": 0.36969452537596226,
+      "reward_std": 0.7297583511099219,
+      "rewards/concensus_correctness_reward_func": 0.056062500923871994,
+      "rewards/consensus_reward_func": 0.0625,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.75,
-      "rewards/question_recreation_reward_func": 0.45432631112635136,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.24291324615478516,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.30899999663233757,
+      "rewards/xmlcount_reward_func": 0.008218752220273018,
       "step": 6
     },
     {
-      "completion_length": 358.90625,
-      "epoch": 1.6,
-      "grad_norm": 2.4348974227905273,
-      "kl": 0.0030339954828377813,
+      "completion_length": 378.9583333333333,
+      "epoch": 2.0,
+      "grad_norm": 17.58197021484375,
+      "kl": 0.0063056297464451445,
       "learning_rate": 3.5042385616324236e-07,
       "loss": 0.0,
-      "reward": 1.108023013919592,
-      "reward_std": 1.2845596000552177,
-      "rewards/concensus_correctness_reward_func": 0.03162499889731407,
-      "rewards/consensus_reward_func": 0.0,
+      "reward": 1.823668549458186,
+      "reward_std": 2.0832131765782833,
+      "rewards/concensus_correctness_reward_func": 0.8333333333333334,
+      "rewards/consensus_reward_func": 0.08333333333333333,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.75,
-      "rewards/question_recreation_reward_func": 0.422929210588336,
+      "rewards/final_correctness_reward_func": 0.4166666666666667,
+      "rewards/question_recreation_reward_func": 0.3380434938396017,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.09653124236501753,
+      "rewards/xmlcount_reward_func": 0.15229166112840176,
       "step": 8
     },
     {
-      "completion_length": 299.84375,
-      "epoch": 2.0,
-      "grad_norm": 2.150526523590088,
-      "kl": 0.0025817730638664216,
+      "completion_length": 319.0,
+      "epoch": 2.571428571428571,
+      "grad_norm": 4.946495056152344,
+      "kl": 0.0023342472850345075,
       "learning_rate": 2.706448363680831e-07,
       "loss": 0.0,
-      "reward": 1.568983219563961,
-      "reward_std": 1.6357312947511673,
-      "rewards/concensus_correctness_reward_func": 0.28462500125169754,
-      "rewards/consensus_reward_func": 0.1875,
+      "reward": 0.6347062969580293,
+      "reward_std": 0.7983217537403107,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0625,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.375,
-      "rewards/question_recreation_reward_func": 0.4329832140356302,
+      "rewards/final_correctness_reward_func": 0.125,
+      "rewards/question_recreation_reward_func": 0.3073625322431326,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.28887500427663326,
+      "rewards/xmlcount_reward_func": 0.1398437414318323,
       "step": 10
     },
     {
-      "completion_length": 389.375,
-      "epoch": 2.4,
-      "grad_norm": 5.240199565887451,
-      "kl": 0.0031134470918914303,
+      "completion_length": 299.9583333333333,
+      "epoch": 3.0,
+      "grad_norm": 32.36935043334961,
+      "kl": 0.002557387478494396,
       "learning_rate": 1.886286282148002e-07,
       "loss": 0.0,
-      "reward": 0.9006264302879572,
-      "reward_std": 1.095049210358411,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.0625,
+      "reward": 1.3059624830881755,
+      "reward_std": 2.1406236762801805,
+      "rewards/concensus_correctness_reward_func": 0.8333333333333334,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.375,
-      "rewards/question_recreation_reward_func": 0.2570951599627733,
+      "rewards/final_correctness_reward_func": 0.3333333333333333,
+      "rewards/question_recreation_reward_func": 0.22862919544180235,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.20603124657645822,
+      "rewards/xmlcount_reward_func": -0.08933333307504654,
       "step": 12
     },
     {
-      "completion_length": 266.34375,
-      "epoch": 2.8,
-      "grad_norm": 2.4637773036956787,
-      "kl": 0.01565292132727336,
+      "completion_length": 364.59375,
+      "epoch": 3.571428571428571,
+      "grad_norm": 6.134871006011963,
+      "kl": 0.002294319980137516,
       "learning_rate": 1.1326296046939333e-07,
       "loss": 0.0,
-      "reward": 1.15899146348238,
-      "reward_std": 1.3713350892066956,
-      "rewards/concensus_correctness_reward_func": 0.05287500098347664,
-      "rewards/consensus_reward_func": 0.125,
+      "reward": 0.9578792434185743,
+      "reward_std": 0.7915452528977767,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0625,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.4375,
-      "rewards/question_recreation_reward_func": 0.39949146285653114,
-      "rewards/soft_format_reward_func": 0.015625,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.12850000336766243,
+      "rewards/final_correctness_reward_func": 0.3125,
+      "rewards/question_recreation_reward_func": 0.3451604973524809,
+      "rewards/soft_format_reward_func": 0.0,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.22209375258535147,
       "step": 14
     },
     {
-      "completion_length": 410.9375,
-      "epoch": 3.2,
-      "grad_norm": 2.6385397911071777,
-      "kl": 0.0014534542278852314,
+      "completion_length": 402.0,
+      "epoch": 4.0,
+      "grad_norm": 1.2527885437011719,
+      "kl": 0.004929080818934987,
       "learning_rate": 5.271487265090163e-08,
       "loss": 0.0,
-      "reward": 0.9041049107909203,
-      "reward_std": 1.032565638422966,
-      "rewards/concensus_correctness_reward_func": 0.03162499889731407,
+      "reward": 0.3584599271416664,
+      "reward_std": 1.0923715432484944,
+      "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.4375,
-      "rewards/question_recreation_reward_func": 0.3778861539904028,
+      "rewards/final_correctness_reward_func": 0.16666666666666666,
+      "rewards/question_recreation_reward_func": 0.2636265928546588,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.05709375347942114,
+      "rewards/xmlcount_reward_func": -0.07183333983023961,
       "step": 16
     },
     {
-      "completion_length": 367.0,
-      "epoch": 3.6,
-      "grad_norm": 1.9641821384429932,
-      "kl": 0.00394120704731904,
+      "completion_length": 376.75,
+      "epoch": 4.571428571428571,
+      "grad_norm": 2.3929598331451416,
+      "kl": 0.002563515809015371,
       "learning_rate": 1.3545689574841341e-08,
       "loss": 0.0,
-      "reward": 0.9352770112454891,
-      "reward_std": 1.2389301266521215,
+      "reward": 0.241681374842301,
+      "reward_std": 1.120352853089571,
       "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.75,
-      "rewards/question_recreation_reward_func": 0.3228082088753581,
+      "rewards/final_correctness_reward_func": 0.1875,
+      "rewards/question_recreation_reward_func": 0.23099386505782604,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.1375312478048727,
+      "rewards/xmlcount_reward_func": -0.1768124857917428,
       "step": 18
     },
     {
-      "completion_length": 319.5,
-      "epoch": 4.0,
-      "grad_norm": 4.13883113861084,
-      "kl": 0.0039633114138268866,
+      "completion_length": 389.0,
+      "epoch": 5.0,
+      "grad_norm": 1.3994996547698975,
+      "kl": 0.002270005274719248,
       "learning_rate": 0.0,
       "loss": 0.0,
-      "reward": 1.2207802087068558,
-      "reward_std": 1.2361758640035987,
-      "rewards/concensus_correctness_reward_func": 0.16231250390410423,
-      "rewards/consensus_reward_func": 0.0625,
+      "reward": 0.5229276654620966,
+      "reward_std": 0.7810152769088745,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.25,
-      "rewards/question_recreation_reward_func": 0.4265302084386349,
+      "rewards/final_correctness_reward_func": 0.16666666666666666,
+      "rewards/question_recreation_reward_func": 0.30830267692605656,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.3038124982267618,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.047958330561717354,
       "step": 20
     },
     {
-      "epoch": 4.0,
+      "epoch": 5.0,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 5.182587085528212e-06,
-      "train_runtime": 867.121,
-      "train_samples_per_second": 0.369,
-      "train_steps_per_second": 0.023
+      "train_loss": 2.5685094556138212e-06,
+      "train_runtime": 536.6447,
+      "train_samples_per_second": 0.596,
+      "train_steps_per_second": 0.037
     }
   ],
   "logging_steps": 2,
   "max_steps": 20,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 4,
+  "num_train_epochs": 7,
   "save_steps": 25,
   "stateful_callbacks": {
     "TrainerControl": {