End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ae1b66e7f7dbbf51da2aaf30c533937df45105662a9cb67102aec0816c53f307
+oid sha256:43cea64e60e2427c950537a34584f5d063e9b9c1b91f67c4e5445a6a293733aa
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 1.1895783154614037e-05,
-    "train_runtime": 436.9213,
+    "train_loss": 1.1039859509764938e-05,
+    "train_runtime": 472.5989,
     "train_samples": 14,
-    "train_samples_per_second": 0.732,
-    "train_steps_per_second": 0.046
+    "train_samples_per_second": 0.677,
+    "train_steps_per_second": 0.042
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 1.1895783154614037e-05,
-    "train_runtime": 436.9213,
+    "train_loss": 1.1039859509764938e-05,
+    "train_runtime": 472.5989,
     "train_samples": 14,
-    "train_samples_per_second": 0.732,
-    "train_steps_per_second": 0.046
+    "train_samples_per_second": 0.677,
+    "train_steps_per_second": 0.042
 }

trainer_state.json

@@ -10,203 +10,203 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 250.25,
+      "completion_length": 297.59375,
       "epoch": 0.5714285714285714,
-      "grad_norm": 1.2837601900100708,
-      "kl": 0.012850156519562006,
+      "grad_norm": 0.5434951186180115,
+      "kl": 0.012450387119315565,
       "learning_rate": 5e-07,
       "loss": 0.0,
-      "reward": 6.77421835064888,
-      "reward_std": 8.612250968813896,
-      "rewards/concensus_correctness_reward_func": 4.25950001180172,
-      "rewards/consensus_reward_func": 0.375,
+      "reward": 3.4174553006887436,
+      "reward_std": 3.5753641426563263,
+      "rewards/concensus_correctness_reward_func": 1.6319999638944864,
+      "rewards/consensus_reward_func": 0.5,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 1.25,
-      "rewards/question_recreation_reward_func": 0.43065583519637585,
+      "rewards/final_correctness_reward_func": 0.25,
+      "rewards/question_recreation_reward_func": 0.5722679821774364,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.4590624952688813,
+      "rewards/xmlcount_reward_func": 0.46318749710917473,
       "step": 2
     },
     {
-      "completion_length": 222.29166666666666,
+      "completion_length": 324.0,
       "epoch": 1.0,
-      "grad_norm": 0.6248174905776978,
-      "kl": 0.013460155421247086,
+      "grad_norm": 0.5409206748008728,
+      "kl": 0.012902182837327322,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 9.162418603897095,
-      "reward_std": 8.077639992038408,
-      "rewards/concensus_correctness_reward_func": 5.69091667731603,
-      "rewards/consensus_reward_func": 0.75,
+      "reward": 8.837965329488119,
+      "reward_std": 7.537450035413106,
+      "rewards/concensus_correctness_reward_func": 5.754750003417333,
+      "rewards/consensus_reward_func": 0.5833333333333334,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 1.4166666666666667,
-      "rewards/question_recreation_reward_func": 0.5355017731587092,
+      "rewards/question_recreation_reward_func": 0.6165488610665003,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.020833333333333332,
-      "rewards/xmlcount_reward_func": 0.7485000193119049,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.46666666865348816,
       "step": 4
     },
     {
-      "completion_length": 277.46875,
+      "completion_length": 292.53125,
       "epoch": 1.5714285714285714,
-      "grad_norm": 0.7902827262878418,
-      "kl": 0.012329760298598558,
+      "grad_norm": 0.6310792565345764,
+      "kl": 0.009227392438333482,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 6.865463137626648,
-      "reward_std": 6.320934921503067,
-      "rewards/concensus_correctness_reward_func": 4.186562471091747,
-      "rewards/consensus_reward_func": 0.375,
+      "reward": 5.818621024489403,
+      "reward_std": 6.194779992103577,
+      "rewards/concensus_correctness_reward_func": 3.9209374338388443,
+      "rewards/consensus_reward_func": 0.5,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 1.375,
-      "rewards/question_recreation_reward_func": 0.4482444394379854,
+      "rewards/final_correctness_reward_func": 0.5,
+      "rewards/question_recreation_reward_func": 0.4342148583382368,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.48065625317394733,
+      "rewards/xmlcount_reward_func": 0.46346874348819256,
       "step": 6
     },
     {
-      "completion_length": 226.625,
+      "completion_length": 229.41666666666666,
       "epoch": 2.0,
-      "grad_norm": 0.5743995308876038,
-      "kl": 0.017415982438251376,
+      "grad_norm": 0.5793716311454773,
+      "kl": 0.016171909558276337,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 8.300778865814209,
-      "reward_std": 7.356443564097087,
-      "rewards/concensus_correctness_reward_func": 5.132833272218704,
-      "rewards/consensus_reward_func": 0.8333333333333334,
+      "reward": 12.946766088406244,
+      "reward_std": 6.9684008955955505,
+      "rewards/concensus_correctness_reward_func": 9.640666643778482,
+      "rewards/consensus_reward_func": 0.9166666666666666,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 1.3333333333333333,
-      "rewards/question_recreation_reward_func": 0.43248720094561577,
+      "rewards/final_correctness_reward_func": 1.1666666666666667,
+      "rewards/question_recreation_reward_func": 0.5147663528720537,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.5687916874885559,
+      "rewards/strict_format_reward_func": 0.041666666666666664,
+      "rewards/xmlcount_reward_func": 0.6663333475589752,
       "step": 8
     },
     {
-      "completion_length": 233.84375,
+      "completion_length": 234.375,
       "epoch": 2.571428571428571,
-      "grad_norm": 0.710727870464325,
-      "kl": 0.011741695983801037,
+      "grad_norm": 0.656565248966217,
+      "kl": 0.014423743879888207,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 6.58307883143425,
-      "reward_std": 6.25597608089447,
-      "rewards/concensus_correctness_reward_func": 3.962499998509884,
-      "rewards/consensus_reward_func": 0.5,
+      "reward": 4.937858000397682,
+      "reward_std": 3.5230171717703342,
+      "rewards/concensus_correctness_reward_func": 2.8464375138282776,
+      "rewards/consensus_reward_func": 0.375,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 1.0625,
-      "rewards/question_recreation_reward_func": 0.46920393593609333,
+      "rewards/final_correctness_reward_func": 0.625,
+      "rewards/question_recreation_reward_func": 0.48098299466073513,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.5888750068843365,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.5948125012218952,
       "step": 10
     },
     {
-      "completion_length": 253.29166666666666,
+      "completion_length": 316.0833333333333,
       "epoch": 3.0,
-      "grad_norm": 0.4581153690814972,
-      "kl": 0.014419077197089791,
+      "grad_norm": 0.43157583475112915,
+      "kl": 0.010882371570914984,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": 8.34391681353251,
-      "reward_std": 5.618081798156102,
-      "rewards/concensus_correctness_reward_func": 5.305249979098638,
+      "reward": 6.302018582820892,
+      "reward_std": 6.069982632994652,
+      "rewards/concensus_correctness_reward_func": 4.324666668971379,
       "rewards/consensus_reward_func": 0.5,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 1.5833333333333333,
-      "rewards/question_recreation_reward_func": 0.40370841696858406,
+      "rewards/final_correctness_reward_func": 0.6666666666666666,
+      "rewards/question_recreation_reward_func": 0.4433935806155205,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.020833333333333332,
-      "rewards/xmlcount_reward_func": 0.5307916762928168,
+      "rewards/xmlcount_reward_func": 0.3464583531022072,
       "step": 12
     },
     {
-      "completion_length": 247.78125,
+      "completion_length": 287.03125,
       "epoch": 3.571428571428571,
-      "grad_norm": 0.8578507304191589,
-      "kl": 0.013678431627340615,
+      "grad_norm": 0.8027998805046082,
+      "kl": 0.012334612431004643,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 6.2360208332538605,
-      "reward_std": 6.665688931941986,
-      "rewards/concensus_correctness_reward_func": 3.136249966919422,
+      "reward": 8.02209883928299,
+      "reward_std": 5.8569860234856606,
+      "rewards/concensus_correctness_reward_func": 5.283124998211861,
       "rewards/consensus_reward_func": 0.5625,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 1.3125,
-      "rewards/question_recreation_reward_func": 0.6277082115411758,
+      "rewards/final_correctness_reward_func": 0.9375,
+      "rewards/question_recreation_reward_func": 0.6241299323737621,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.5970625095069408,
+      "rewards/xmlcount_reward_func": 0.6148437671363354,
       "step": 14
     },
     {
-      "completion_length": 293.2083333333333,
+      "completion_length": 304.5,
       "epoch": 4.0,
-      "grad_norm": 0.43180230259895325,
-      "kl": 0.012424148308734098,
+      "grad_norm": 0.597561776638031,
+      "kl": 0.011826164244363705,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 7.924095153808594,
-      "reward_std": 11.039137999216715,
-      "rewards/concensus_correctness_reward_func": 5.284583330154419,
-      "rewards/consensus_reward_func": 0.4166666666666667,
+      "reward": 5.414454311132431,
+      "reward_std": 7.721007605393727,
+      "rewards/concensus_correctness_reward_func": 3.5014999906222024,
+      "rewards/consensus_reward_func": 0.3333333333333333,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 1.5833333333333333,
-      "rewards/question_recreation_reward_func": 0.3320951734979947,
+      "rewards/final_correctness_reward_func": 1.0,
+      "rewards/question_recreation_reward_func": 0.3357460598150889,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.3074166687826316,
+      "rewards/strict_format_reward_func": 0.020833333333333332,
+      "rewards/xmlcount_reward_func": 0.2230416772266229,
       "step": 16
     },
     {
-      "completion_length": 244.5,
+      "completion_length": 284.4375,
       "epoch": 4.571428571428571,
-      "grad_norm": 1.2633910179138184,
-      "kl": 0.014250321430154145,
+      "grad_norm": 0.6247209310531616,
+      "kl": 0.013364863698370755,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 4.794973254203796,
-      "reward_std": 4.659415230154991,
-      "rewards/concensus_correctness_reward_func": 2.2643124759197235,
-      "rewards/consensus_reward_func": 0.1875,
+      "reward": 5.225602433085442,
+      "reward_std": 5.357156628742814,
+      "rewards/concensus_correctness_reward_func": 2.7117499969899654,
+      "rewards/consensus_reward_func": 0.6875,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 1.25,
-      "rewards/question_recreation_reward_func": 0.5774108543992043,
+      "rewards/final_correctness_reward_func": 0.6875,
+      "rewards/question_recreation_reward_func": 0.6408212631940842,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.500124990940094,
+      "rewards/xmlcount_reward_func": 0.48240625113248825,
       "step": 18
     },
     {
-      "completion_length": 235.125,
+      "completion_length": 247.79166666666666,
       "epoch": 5.0,
-      "grad_norm": 0.6901851296424866,
-      "kl": 0.01442628400400281,
+      "grad_norm": 0.8443843722343445,
+      "kl": 0.013195230159908533,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 5.657399932543437,
-      "reward_std": 3.9982150395711265,
-      "rewards/concensus_correctness_reward_func": 2.6396666218837104,
-      "rewards/consensus_reward_func": 0.8333333333333334,
+      "reward": 10.117159684499105,
+      "reward_std": 8.107400243481,
+      "rewards/concensus_correctness_reward_func": 7.329333300391833,
+      "rewards/consensus_reward_func": 0.75,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 1.4166666666666667,
-      "rewards/question_recreation_reward_func": 0.29173319910963375,
-      "rewards/soft_format_reward_func": 0.0,
+      "rewards/final_correctness_reward_func": 1.1666666666666667,
+      "rewards/question_recreation_reward_func": 0.47482667366663617,
+      "rewards/soft_format_reward_func": 0.020833333333333332,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.4760000060002009,
+      "rewards/xmlcount_reward_func": 0.3755000059803327,
       "step": 20
     },
     {
       "epoch": 5.0,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 1.1895783154614037e-05,
-      "train_runtime": 436.9213,
-      "train_samples_per_second": 0.732,
-      "train_steps_per_second": 0.046
+      "train_loss": 1.1039859509764938e-05,
+      "train_runtime": 472.5989,
+      "train_samples_per_second": 0.677,
+      "train_steps_per_second": 0.042
     }
   ],
   "logging_steps": 2,