End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:06a74e6e2d014916f90ee690dce3c50c22bdd08d43b69654327d11f78b7067b2
+oid sha256:1d54d78c6f4c4790d352d2a528e3cc7b2e587c898b776eb0eb058eeb49eda159
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 3.4546391179901545e-05,
-    "train_runtime": 941.4145,
-    "train_samples": 15,
-    "train_samples_per_second": 0.34,
-    "train_steps_per_second": 0.021
+    "train_loss": 3.433250567468349e-05,
+    "train_runtime": 989.3924,
+    "train_samples": 16,
+    "train_samples_per_second": 0.323,
+    "train_steps_per_second": 0.02
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 3.4546391179901545e-05,
-    "train_runtime": 941.4145,
-    "train_samples": 15,
-    "train_samples_per_second": 0.34,
-    "train_steps_per_second": 0.021
+    "train_loss": 3.433250567468349e-05,
+    "train_runtime": 989.3924,
+    "train_samples": 16,
+    "train_samples_per_second": 0.323,
+    "train_steps_per_second": 0.02
 }

trainer_state.json

@@ -10,209 +10,209 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 198.40625,
-      "epoch": 0.5333333333333333,
-      "grad_norm": 8.360007286071777,
-      "kl": 0.03296396473888308,
+      "completion_length": 259.4375,
+      "epoch": 0.5,
+      "grad_norm": 0.8540968894958496,
+      "kl": 0.03679677110631019,
       "learning_rate": 5e-07,
       "loss": 0.0,
-      "reward": 6.321916818618774,
-      "reward_std": 1.7188230901956558,
-      "rewards/concensus_correctness_reward_func": 2.3461874574422836,
-      "rewards/consensus_reward_func": 1.5625,
+      "reward": 4.937266618013382,
+      "reward_std": 2.7515269592404366,
+      "rewards/concensus_correctness_reward_func": 1.429187498986721,
+      "rewards/consensus_reward_func": 1.0,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.875,
-      "rewards/question_recreation_reward_func": 0.48694803565740585,
+      "rewards/question_recreation_reward_func": 0.7472665458917618,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0625,
-      "rewards/xmlcount_reward_func": 0.9887812659144402,
+      "rewards/strict_format_reward_func": 0.046875,
+      "rewards/xmlcount_reward_func": 0.8389374949038029,
       "step": 2
     },
     {
-      "completion_length": 228.32142857142858,
+      "completion_length": 227.90625,
       "epoch": 1.0,
-      "grad_norm": 0.6417061686515808,
-      "kl": 0.029710494380976473,
+      "grad_norm": 1.4667596817016602,
+      "kl": 0.03782109543681145,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 5.377460071018764,
-      "reward_std": 1.0338971018791199,
-      "rewards/concensus_correctness_reward_func": 1.8204285672732763,
-      "rewards/consensus_reward_func": 1.3571428571428572,
+      "reward": 6.107743978500366,
+      "reward_std": 2.6981655061244965,
+      "rewards/concensus_correctness_reward_func": 2.093937508761883,
+      "rewards/consensus_reward_func": 1.5,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.7857142857142857,
-      "rewards/question_recreation_reward_func": 0.5691385694912502,
+      "rewards/final_correctness_reward_func": 1.0625,
+      "rewards/question_recreation_reward_func": 0.6069939360022545,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.03571428571428571,
-      "rewards/xmlcount_reward_func": 0.8093214460781643,
+      "rewards/strict_format_reward_func": 0.046875,
+      "rewards/xmlcount_reward_func": 0.7974375076591969,
       "step": 4
     },
     {
-      "completion_length": 184.96875,
-      "epoch": 1.5333333333333332,
-      "grad_norm": 1.40841543674469,
-      "kl": 0.04421051638200879,
+      "completion_length": 278.875,
+      "epoch": 1.5,
+      "grad_norm": 1.112252116203308,
+      "kl": 0.02613301109522581,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 5.1417646408081055,
-      "reward_std": 2.2220220267772675,
-      "rewards/concensus_correctness_reward_func": 1.5205624923110008,
-      "rewards/consensus_reward_func": 1.25,
+      "reward": 4.020417332649231,
+      "reward_std": 2.134015291929245,
+      "rewards/concensus_correctness_reward_func": 1.2185000143945217,
+      "rewards/consensus_reward_func": 0.9375,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.6875,
-      "rewards/question_recreation_reward_func": 0.4983896389603615,
+      "rewards/final_correctness_reward_func": 0.75,
+      "rewards/question_recreation_reward_func": 0.5292610581964254,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.125,
-      "rewards/xmlcount_reward_func": 1.060312531888485,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.5851562507450581,
       "step": 6
     },
     {
-      "completion_length": 243.64285714285714,
+      "completion_length": 196.25,
       "epoch": 2.0,
-      "grad_norm": 0.5562111735343933,
-      "kl": 0.02644094345825059,
+      "grad_norm": 1.91428804397583,
+      "kl": 0.03515635652001947,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 5.550810728754316,
-      "reward_std": 1.8111653668539864,
-      "rewards/concensus_correctness_reward_func": 1.8958571127482824,
-      "rewards/consensus_reward_func": 1.2142857142857142,
+      "reward": 5.238867521286011,
+      "reward_std": 1.8473113477230072,
+      "rewards/concensus_correctness_reward_func": 1.5862500071525574,
+      "rewards/consensus_reward_func": 1.125,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 1.0714285714285714,
-      "rewards/question_recreation_reward_func": 0.5573822898524148,
-      "rewards/soft_format_reward_func": 0.017857142857142856,
-      "rewards/strict_format_reward_func": 0.05357142857142857,
-      "rewards/xmlcount_reward_func": 0.7404285839625767,
+      "rewards/final_correctness_reward_func": 1.125,
+      "rewards/question_recreation_reward_func": 0.5618673153221607,
+      "rewards/soft_format_reward_func": 0.0,
+      "rewards/strict_format_reward_func": 0.03125,
+      "rewards/xmlcount_reward_func": 0.8095000050961971,
       "step": 8
     },
     {
-      "completion_length": 186.96875,
-      "epoch": 2.533333333333333,
-      "grad_norm": 0.8356194496154785,
-      "kl": 0.03581268759444356,
+      "completion_length": 232.78125,
+      "epoch": 2.5,
+      "grad_norm": 0.8493509292602539,
+      "kl": 0.034787516575306654,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 4.26269268989563,
-      "reward_std": 1.9406834915280342,
-      "rewards/concensus_correctness_reward_func": 1.3394375033676624,
-      "rewards/consensus_reward_func": 1.0,
+      "reward": 4.2884341776371,
+      "reward_std": 2.0309074632823467,
+      "rewards/concensus_correctness_reward_func": 1.0654374957084656,
+      "rewards/consensus_reward_func": 0.875,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.4375,
-      "rewards/question_recreation_reward_func": 0.5125052509829402,
-      "rewards/soft_format_reward_func": 0.015625,
-      "rewards/strict_format_reward_func": 0.09375,
-      "rewards/xmlcount_reward_func": 0.8638750165700912,
+      "rewards/final_correctness_reward_func": 0.9375,
+      "rewards/question_recreation_reward_func": 0.6273716315627098,
+      "rewards/soft_format_reward_func": 0.0,
+      "rewards/strict_format_reward_func": 0.0625,
+      "rewards/xmlcount_reward_func": 0.7206249912269413,
       "step": 10
     },
     {
-      "completion_length": 207.5,
+      "completion_length": 216.84375,
       "epoch": 3.0,
-      "grad_norm": 0.9257772564888,
-      "kl": 0.038460450513022285,
+      "grad_norm": 1.0730605125427246,
+      "kl": 0.040554615668952465,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": 6.022549152374268,
-      "reward_std": 2.5043100884982517,
-      "rewards/concensus_correctness_reward_func": 2.333285689353943,
-      "rewards/consensus_reward_func": 1.2857142857142858,
+      "reward": 4.705726832151413,
+      "reward_std": 1.8564081564545631,
+      "rewards/concensus_correctness_reward_func": 1.1792500019073486,
+      "rewards/consensus_reward_func": 1.0625,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.7857142857142857,
-      "rewards/question_recreation_reward_func": 0.4640133615050997,
+      "rewards/final_correctness_reward_func": 1.0625,
+      "rewards/question_recreation_reward_func": 0.40166423469781876,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.14285714285714285,
-      "rewards/xmlcount_reward_func": 1.0109642914363317,
+      "rewards/strict_format_reward_func": 0.0625,
+      "rewards/xmlcount_reward_func": 0.9373124986886978,
       "step": 12
     },
     {
-      "completion_length": 193.4375,
-      "epoch": 3.533333333333333,
-      "grad_norm": 0.8213157057762146,
-      "kl": 0.046339198597706854,
+      "completion_length": 287.96875,
+      "epoch": 3.5,
+      "grad_norm": 1.0202021598815918,
+      "kl": 0.02543360088020563,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 4.843554437160492,
-      "reward_std": 1.867899501696229,
-      "rewards/concensus_correctness_reward_func": 1.601437471807003,
+      "reward": 5.909228906035423,
+      "reward_std": 3.1725671738386154,
+      "rewards/concensus_correctness_reward_func": 2.2015625089406967,
       "rewards/consensus_reward_func": 1.125,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.625,
-      "rewards/question_recreation_reward_func": 0.4974917620420456,
+      "rewards/final_correctness_reward_func": 1.1875,
+      "rewards/question_recreation_reward_func": 0.565166411921382,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.09375,
-      "rewards/xmlcount_reward_func": 0.9008749946951866,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.8143750131130219,
       "step": 14
     },
     {
-      "completion_length": 214.35714285714286,
+      "completion_length": 203.34375,
       "epoch": 4.0,
-      "grad_norm": 0.7088996171951294,
-      "kl": 0.04368409647473267,
+      "grad_norm": 1.7447508573532104,
+      "kl": 0.04108604905195534,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 5.192878893443516,
-      "reward_std": 2.4513211250305176,
-      "rewards/concensus_correctness_reward_func": 1.527428584439414,
-      "rewards/consensus_reward_func": 1.0,
+      "reward": 4.100949555635452,
+      "reward_std": 1.6319950204342604,
+      "rewards/concensus_correctness_reward_func": 1.246437482535839,
+      "rewards/consensus_reward_func": 1.0625,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 1.1428571428571428,
-      "rewards/question_recreation_reward_func": 0.4679503919822829,
+      "rewards/final_correctness_reward_func": 0.5625,
+      "rewards/question_recreation_reward_func": 0.3207308240234852,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.10714285714285714,
-      "rewards/xmlcount_reward_func": 0.9475000245230538,
+      "rewards/strict_format_reward_func": 0.0625,
+      "rewards/xmlcount_reward_func": 0.8462812379002571,
       "step": 16
     },
     {
-      "completion_length": 200.65625,
-      "epoch": 4.533333333333333,
-      "grad_norm": 0.7495886087417603,
-      "kl": 0.027507278718985617,
+      "completion_length": 224.03125,
+      "epoch": 4.5,
+      "grad_norm": 1.173526644706726,
+      "kl": 0.03312400449067354,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 6.123600482940674,
-      "reward_std": 1.2783842608332634,
-      "rewards/concensus_correctness_reward_func": 2.145937453955412,
-      "rewards/consensus_reward_func": 1.5,
+      "reward": 5.5468989610672,
+      "reward_std": 2.411236487329006,
+      "rewards/concensus_correctness_reward_func": 1.8100625053048134,
+      "rewards/consensus_reward_func": 1.3125,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.875,
-      "rewards/question_recreation_reward_func": 0.5176317282021046,
+      "rewards/final_correctness_reward_func": 1.0,
+      "rewards/question_recreation_reward_func": 0.43858631886541843,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0625,
-      "rewards/xmlcount_reward_func": 1.0225312784314156,
+      "rewards/strict_format_reward_func": 0.09375,
+      "rewards/xmlcount_reward_func": 0.8919999897480011,
       "step": 18
     },
     {
-      "completion_length": 201.32142857142858,
+      "completion_length": 244.09375,
       "epoch": 5.0,
-      "grad_norm": 0.6576955914497375,
-      "kl": 0.04337019233831337,
+      "grad_norm": 1.314996600151062,
+      "kl": 0.031924105947837234,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 6.447563511984689,
-      "reward_std": 1.7632163337298803,
-      "rewards/concensus_correctness_reward_func": 2.3713571344103133,
-      "rewards/consensus_reward_func": 1.7142857142857142,
+      "reward": 3.978771150112152,
+      "reward_std": 1.9551344588398933,
+      "rewards/concensus_correctness_reward_func": 1.0894374959170818,
+      "rewards/consensus_reward_func": 0.75,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 1.0,
-      "rewards/question_recreation_reward_func": 0.4367780275642872,
+      "rewards/final_correctness_reward_func": 0.625,
+      "rewards/question_recreation_reward_func": 0.5835523679852486,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.08928571428571429,
-      "rewards/xmlcount_reward_func": 0.8358571614537921,
+      "rewards/strict_format_reward_func": 0.078125,
+      "rewards/xmlcount_reward_func": 0.852656252682209,
       "step": 20
     },
     {
       "epoch": 5.0,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 3.4546391179901545e-05,
-      "train_runtime": 941.4145,
-      "train_samples_per_second": 0.34,
-      "train_steps_per_second": 0.021
+      "train_loss": 3.433250567468349e-05,
+      "train_runtime": 989.3924,
+      "train_samples_per_second": 0.323,
+      "train_steps_per_second": 0.02
     }
   ],
   "logging_steps": 2,
   "max_steps": 20,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 7,
+  "num_train_epochs": 5,
   "save_steps": 25,
   "stateful_callbacks": {
     "TrainerControl": {