End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c2f6ce216bf587af85e89f212099466d2699f5f0256ff7a22350016c50459b30
+oid sha256:afd912389d6f8c8e8da616bd6063b6c09dd9837ced1bebcc45b421b5f3193cb7
 size 35237104

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 5.493214257512591e-06,
-    "train_runtime": 1348.1291,
+    "train_loss": 4.3360367385503196e-06,
+    "train_runtime": 977.7444,
     "train_samples": 11,
-    "train_samples_per_second": 0.237,
-    "train_steps_per_second": 0.015
+    "train_samples_per_second": 0.327,
+    "train_steps_per_second": 0.02
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 5.493214257512591e-06,
-    "train_runtime": 1348.1291,
+    "train_loss": 4.3360367385503196e-06,
+    "train_runtime": 977.7444,
     "train_samples": 11,
-    "train_samples_per_second": 0.237,
-    "train_steps_per_second": 0.015
+    "train_samples_per_second": 0.327,
+    "train_steps_per_second": 0.02
 }

trainer_state.json

@@ -10,203 +10,203 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 309.8125,
+      "completion_length": 276.46875,
       "epoch": 0.7272727272727273,
-      "grad_norm": 8.192431449890137,
-      "kl": 0.0013196841391618364,
+      "grad_norm": 1.8341317176818848,
+      "kl": 0.014543038392730523,
       "learning_rate": 4.965903258506806e-07,
       "loss": 0.0,
-      "reward": 0.22969480406027287,
-      "reward_std": 0.39862296264618635,
+      "reward": 0.21561870630830526,
+      "reward_std": 0.18531930854078382,
       "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.125,
-      "rewards/question_recreation_reward_func": 0.09497605240903795,
+      "rewards/final_correctness_reward_func": 0.0625,
+      "rewards/question_recreation_reward_func": 0.07921245764009655,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.00971875200048089,
+      "rewards/xmlcount_reward_func": 0.07390624936670065,
       "step": 2
     },
     {
-      "completion_length": 334.5,
+      "completion_length": 266.25,
       "epoch": 1.3636363636363638,
-      "grad_norm": 16.709056854248047,
-      "kl": 0.01644311550106587,
+      "grad_norm": 37.67310333251953,
+      "kl": 0.0022699503793514202,
       "learning_rate": 4.698684378016222e-07,
       "loss": 0.0,
-      "reward": 1.084287445460047,
-      "reward_std": 1.6699770393648319,
+      "reward": 0.7655775014843259,
+      "reward_std": 1.7434283659926482,
       "rewards/concensus_correctness_reward_func": 0.7142857142857143,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.14285714285714285,
-      "rewards/question_recreation_reward_func": 0.17485884497208254,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.1444345811115844,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.05228571646979877,
+      "rewards/xmlcount_reward_func": -0.09314285218715668,
       "step": 4
     },
     {
-      "completion_length": 345.17857142857144,
+      "completion_length": 329.7857142857143,
       "epoch": 2.0,
-      "grad_norm": 2.4254114627838135,
-      "kl": 0.001126350327727518,
+      "grad_norm": 5.340218544006348,
+      "kl": 0.0013989850163592824,
       "learning_rate": 4.193203929064353e-07,
       "loss": 0.0,
-      "reward": 0.16946781054139137,
-      "reward_std": 0.30478086562028955,
+      "reward": 0.24973871931433678,
+      "reward_std": 0.2699835923101221,
       "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.1678249617772443,
+      "rewards/question_recreation_reward_func": 0.18173872173896857,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.0016428555494972638,
+      "rewards/xmlcount_reward_func": 0.06799999943801335,
       "step": 6
     },
     {
-      "completion_length": 426.03125,
+      "completion_length": 367.8125,
       "epoch": 2.7272727272727275,
-      "grad_norm": 1.3110556602478027,
-      "kl": 0.0015403831857838668,
+      "grad_norm": 4.607388973236084,
+      "kl": 0.002823736780555919,
       "learning_rate": 3.5042385616324236e-07,
       "loss": 0.0,
-      "reward": 0.26408953219652176,
-      "reward_std": 0.27848068880848587,
+      "reward": 0.15139904618263245,
+      "reward_std": 0.2963482877239585,
       "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.18487078556790948,
+      "rewards/question_recreation_reward_func": 0.13202404603362083,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.07921875081956387,
+      "rewards/xmlcount_reward_func": 0.019375001080334187,
       "step": 8
     },
     {
-      "completion_length": 334.10714285714283,
+      "completion_length": 244.60714285714286,
       "epoch": 3.3636363636363638,
-      "grad_norm": 1.4304771423339844,
-      "kl": 0.0011949114850722253,
+      "grad_norm": 5.428980827331543,
+      "kl": 0.0022493676640026805,
       "learning_rate": 2.706448363680831e-07,
       "loss": 0.0,
-      "reward": 0.2872551094208445,
-      "reward_std": 0.42002125616584507,
+      "reward": 0.3734208771160671,
+      "reward_std": 0.33582321447985514,
       "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.18396939922656333,
+      "rewards/question_recreation_reward_func": 0.25692086826477734,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.1032857107264655,
+      "rewards/xmlcount_reward_func": 0.11649999873978752,
       "step": 10
     },
     {
-      "completion_length": 290.07142857142856,
+      "completion_length": 342.4642857142857,
       "epoch": 4.0,
-      "grad_norm": 35.479034423828125,
-      "kl": 0.020025783434643278,
+      "grad_norm": 44.4582405090332,
+      "kl": 0.007347333985048213,
       "learning_rate": 1.886286282148002e-07,
       "loss": 0.0,
-      "reward": 0.2351770829409361,
-      "reward_std": 0.317613242992333,
+      "reward": 0.09124727666910205,
+      "reward_std": 0.08825840002724103,
       "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.12906994510974204,
+      "rewards/question_recreation_reward_func": 0.07117584853300027,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.10610714235476085,
+      "rewards/xmlcount_reward_func": 0.020071428269147873,
       "step": 12
     },
     {
-      "completion_length": 385.1875,
+      "completion_length": 373.0625,
       "epoch": 4.7272727272727275,
-      "grad_norm": 1.7368310689926147,
-      "kl": 0.0016509672495885752,
+      "grad_norm": 13.767102241516113,
+      "kl": 0.006031289682141505,
       "learning_rate": 1.1326296046939333e-07,
       "loss": 0.0,
-      "reward": 0.14453346747905016,
-      "reward_std": 0.6091575883328915,
+      "reward": 0.29733651620335877,
+      "reward_std": 0.38564108312129974,
       "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": 0.179158469196409,
+      "rewards/final_correctness_reward_func": 0.0625,
+      "rewards/question_recreation_reward_func": 0.12352400948293507,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.2846250347793102,
+      "rewards/xmlcount_reward_func": 0.11131250020116568,
       "step": 14
     },
     {
-      "completion_length": 329.92857142857144,
+      "completion_length": 387.5357142857143,
       "epoch": 5.363636363636363,
-      "grad_norm": 16.084075927734375,
-      "kl": 0.00911175856266969,
+      "grad_norm": 2.061283588409424,
+      "kl": 0.0017532423038833908,
       "learning_rate": 5.271487265090163e-08,
       "loss": 0.0,
-      "reward": 0.08588471157210213,
-      "reward_std": 0.1177412923425436,
+      "reward": 0.07720165061099189,
+      "reward_std": 0.32941546191328336,
       "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.047313284394996505,
+      "rewards/question_recreation_reward_func": 0.1725587903388909,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.038571428773658614,
+      "rewards/xmlcount_reward_func": -0.09535714398537364,
       "step": 16
     },
     {
-      "completion_length": 295.35714285714283,
+      "completion_length": 398.32142857142856,
       "epoch": 6.0,
-      "grad_norm": 27.753393173217773,
-      "kl": 0.006975149270147085,
+      "grad_norm": 1.1391915082931519,
+      "kl": 0.00572152525585677,
       "learning_rate": 1.3545689574841341e-08,
       "loss": 0.0,
-      "reward": 0.4774811161415918,
-      "reward_std": 0.5152423062494823,
+      "reward": 0.15062752525721276,
+      "reward_std": 0.43591643816658426,
       "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.14285714285714285,
-      "rewards/question_recreation_reward_func": 0.15815968705075129,
+      "rewards/final_correctness_reward_func": 0.07142857142857142,
+      "rewards/question_recreation_reward_func": 0.13544895446726254,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.17646428516932897,
+      "rewards/xmlcount_reward_func": -0.056249999574252536,
       "step": 18
     },
     {
-      "completion_length": 244.34375,
+      "completion_length": 377.5,
       "epoch": 6.7272727272727275,
-      "grad_norm": 13.965374946594238,
-      "kl": 0.00237850533449091,
+      "grad_norm": 1.4635015726089478,
+      "kl": 0.0017910118331201375,
       "learning_rate": 0.0,
       "loss": 0.0,
-      "reward": 0.6199594191275537,
-      "reward_std": 0.5588421156862751,
-      "rewards/concensus_correctness_reward_func": 0.09943749941885471,
-      "rewards/consensus_reward_func": 0.0625,
+      "reward": 0.020720995496958494,
+      "reward_std": 0.6026385057484731,
+      "rewards/concensus_correctness_reward_func": 0.0,
+      "rewards/consensus_reward_func": 0.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.125,
-      "rewards/question_recreation_reward_func": 0.19514692039228976,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.1422209832817316,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.137874998152256,
+      "rewards/xmlcount_reward_func": -0.12149999663233757,
       "step": 20
     },
     {
       "epoch": 6.7272727272727275,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 5.493214257512591e-06,
-      "train_runtime": 1348.1291,
-      "train_samples_per_second": 0.237,
-      "train_steps_per_second": 0.015
+      "train_loss": 4.3360367385503196e-06,
+      "train_runtime": 977.7444,
+      "train_samples_per_second": 0.327,
+      "train_steps_per_second": 0.02
     }
   ],
   "logging_steps": 2,