End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1476e33a5ed534f9186b82150074229a5ae8fdc2efa04b620ad2493b10b3c416
+oid sha256:facca0073fa6447f963df4f1448f8a659bff56b9096e7a94e72f4c35db26c5c0
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 7.247691939937795e-07,
-    "train_runtime": 1796.7992,
-    "train_samples": 26,
-    "train_samples_per_second": 0.178,
-    "train_steps_per_second": 0.011
+    "train_loss": 6.874654843613825e-07,
+    "train_runtime": 1539.8882,
+    "train_samples": 11,
+    "train_samples_per_second": 0.208,
+    "train_steps_per_second": 0.013
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 7.247691939937795e-07,
-    "train_runtime": 1796.7992,
-    "train_samples": 26,
-    "train_samples_per_second": 0.178,
-    "train_steps_per_second": 0.011
+    "train_loss": 6.874654843613825e-07,
+    "train_runtime": 1539.8882,
+    "train_samples": 11,
+    "train_samples_per_second": 0.208,
+    "train_steps_per_second": 0.013
 }

trainer_state.json

@@ -2,7 +2,7 @@
   "best_global_step": null,
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 2.9230769230769234,
+  "epoch": 6.7272727272727275,
   "eval_steps": 500,
   "global_step": 20,
   "is_hyper_param_search": false,
@@ -10,209 +10,209 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 411.09375,
-      "epoch": 0.3076923076923077,
-      "grad_norm": 0.6646990776062012,
-      "kl": 0.0008481693439534865,
+      "completion_length": 376.34375,
+      "epoch": 0.7272727272727273,
+      "grad_norm": 2.123795509338379,
+      "kl": 0.0008059284700721037,
       "learning_rate": 5e-07,
       "loss": 0.0,
-      "reward": 0.8132255133241415,
-      "reward_std": 0.9614562764763832,
+      "reward": 0.8194242492318153,
+      "reward_std": 0.8725666627287865,
       "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.1875,
+      "rewards/consensus_reward_func": 0.3125,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5623192805796862,
+      "rewards/question_recreation_reward_func": 0.4507679883390665,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.04778125882148743,
+      "rewards/xmlcount_reward_func": 0.040531257865950465,
       "step": 2
     },
     {
-      "completion_length": 520.9375,
-      "epoch": 0.6153846153846154,
-      "grad_norm": 0.3529391586780548,
-      "kl": 0.0006171913410071284,
+      "completion_length": 463.57142857142856,
+      "epoch": 1.3636363636363638,
+      "grad_norm": 0.444406121969223,
+      "kl": 0.0006558234370978815,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 1.0342297032475471,
-      "reward_std": 1.2277675792574883,
+      "reward": 1.0296695147241866,
+      "reward_std": 0.8633963976587568,
       "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.4375,
+      "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.5535734333097935,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.23541951019849097,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.019343746826052666,
+      "rewards/strict_format_reward_func": 0.03571428571428571,
+      "rewards/xmlcount_reward_func": 0.5442500114440918,
       "step": 4
     },
     {
-      "completion_length": 539.6875,
-      "epoch": 0.9230769230769231,
-      "grad_norm": 0.36468133330345154,
-      "kl": 0.0007185676295193844,
+      "completion_length": 598.5,
+      "epoch": 2.0,
+      "grad_norm": 0.40745630860328674,
+      "kl": 0.0007307843105601412,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 1.0436064377427101,
-      "reward_std": 0.7794338203966618,
+      "reward": 0.7694277209894997,
+      "reward_std": 1.1632391554968697,
       "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.375,
+      "rewards/consensus_reward_func": 0.2857142857142857,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.125,
-      "rewards/question_recreation_reward_func": 0.6074189431965351,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5034991694348199,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.0638125017285347,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": -0.037642872759274075,
       "step": 6
     },
     {
-      "completion_length": 459.7916666666667,
-      "epoch": 1.1538461538461537,
-      "grad_norm": 0.39633116126060486,
-      "kl": 0.0006742304831277579,
+      "completion_length": 560.1875,
+      "epoch": 2.7272727272727275,
+      "grad_norm": 0.2870396077632904,
+      "kl": 0.0006468838182627223,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 1.105679822464784,
-      "reward_std": 1.0789686491092045,
+      "reward": 0.7364773657172918,
+      "reward_std": 1.0139315500855446,
       "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.5,
+      "rewards/consensus_reward_func": 0.25,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.4385964597264926,
+      "rewards/question_recreation_reward_func": 0.3468836098909378,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.020833333333333332,
-      "rewards/xmlcount_reward_func": 0.14624997973442078,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.1239687567576766,
       "step": 8
     },
     {
-      "completion_length": 453.9375,
-      "epoch": 1.4615384615384617,
-      "grad_norm": 0.6026042699813843,
-      "kl": 0.0008009431621758267,
+      "completion_length": 543.0,
+      "epoch": 3.3636363636363638,
+      "grad_norm": 0.4253978729248047,
+      "kl": 0.0007500549545511603,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 0.9355399496853352,
-      "reward_std": 0.8070156015455723,
+      "reward": 0.5847084107143539,
+      "reward_std": 1.0092563756874628,
       "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.25,
+      "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5751336961984634,
+      "rewards/question_recreation_reward_func": 0.42106556040900095,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.09478125628083944,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": -0.06849999725818634,
       "step": 10
     },
     {
-      "completion_length": 486.25,
-      "epoch": 1.7692307692307692,
-      "grad_norm": 0.39988255500793457,
-      "kl": 0.000704013007634785,
+      "completion_length": 533.6428571428571,
+      "epoch": 4.0,
+      "grad_norm": 0.30162712931632996,
+      "kl": 0.0008266470395028591,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": 1.1202685348689556,
-      "reward_std": 0.8506302107125521,
+      "reward": 0.4913223215511867,
+      "reward_std": 0.9674114840371268,
       "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.5,
+      "rewards/consensus_reward_func": 0.07142857142857142,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.4937372952699661,
+      "rewards/question_recreation_reward_func": 0.40028661489486694,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.11090626008808613,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.0017500030142920358,
       "step": 12
     },
     {
-      "completion_length": 463.125,
-      "epoch": 2.0,
-      "grad_norm": 0.23108026385307312,
-      "kl": 0.0007660971896257252,
+      "completion_length": 566.34375,
+      "epoch": 4.7272727272727275,
+      "grad_norm": 0.6862844824790955,
+      "kl": 0.0007115659973351285,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 0.5256565163532892,
-      "reward_std": 1.0281643321116765,
+      "reward": 0.9112295396625996,
+      "reward_std": 1.0470195040106773,
       "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.16666666666666666,
+      "rewards/consensus_reward_func": 0.1875,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.541198194026947,
+      "rewards/question_recreation_reward_func": 0.4396982938051224,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.020833333333333332,
-      "rewards/xmlcount_reward_func": -0.20304166773955026,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.26840626634657383,
       "step": 14
     },
     {
-      "completion_length": 473.5625,
-      "epoch": 2.3076923076923075,
-      "grad_norm": 0.44562846422195435,
-      "kl": 0.0007443879439961165,
+      "completion_length": 535.8571428571429,
+      "epoch": 5.363636363636363,
+      "grad_norm": 0.2605418860912323,
+      "kl": 0.0005985684318667543,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 0.8368261530995369,
-      "reward_std": 0.9730200506746769,
+      "reward": 0.9383800370352608,
+      "reward_std": 0.9764512436730521,
       "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.125,
+      "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.0625,
-      "rewards/question_recreation_reward_func": 0.45257615856826305,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.42352288003478733,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.18112499732524157,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.28271428868174553,
       "step": 16
     },
     {
-      "completion_length": 443.15625,
-      "epoch": 2.6153846153846154,
-      "grad_norm": 0.4637618660926819,
-      "kl": 0.000774655825807713,
+      "completion_length": 533.7857142857143,
+      "epoch": 6.0,
+      "grad_norm": 3.9433951377868652,
+      "kl": 0.0008690485265105963,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 1.0078877992928028,
-      "reward_std": 0.9847271293401718,
+      "reward": 1.395365025315966,
+      "reward_std": 1.3007023589951652,
       "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.3125,
+      "rewards/consensus_reward_func": 0.5,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5618877746164799,
+      "rewards/question_recreation_reward_func": 0.3816864735313824,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.1335000079125166,
+      "rewards/strict_format_reward_func": 0.07142857142857142,
+      "rewards/xmlcount_reward_func": 0.4422500016433852,
       "step": 18
     },
     {
-      "completion_length": 312.65625,
-      "epoch": 2.9230769230769234,
-      "grad_norm": 0.6996157169342041,
-      "kl": 0.0010156768184970133,
+      "completion_length": 471.875,
+      "epoch": 6.7272727272727275,
+      "grad_norm": 0.6051961183547974,
+      "kl": 0.0007186525617726147,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 1.2219690904021263,
-      "reward_std": 0.9160432629287243,
+      "reward": 0.9024831224232912,
+      "reward_std": 1.044289842247963,
       "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.4375,
+      "rewards/consensus_reward_func": 0.25,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.4381253346800804,
+      "rewards/question_recreation_reward_func": 0.4804831203073263,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.046875,
-      "rewards/xmlcount_reward_func": 0.29946876410394907,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.17199999652802944,
       "step": 20
     },
     {
-      "epoch": 2.9230769230769234,
+      "epoch": 6.7272727272727275,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 7.247691939937795e-07,
-      "train_runtime": 1796.7992,
-      "train_samples_per_second": 0.178,
-      "train_steps_per_second": 0.011
+      "train_loss": 6.874654843613825e-07,
+      "train_runtime": 1539.8882,
+      "train_samples_per_second": 0.208,
+      "train_steps_per_second": 0.013
     }
   ],
   "logging_steps": 2,
   "max_steps": 20,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 4,
+  "num_train_epochs": 10,
   "save_steps": 25,
   "stateful_callbacks": {
     "TrainerControl": {