End of training

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8f4f00b6a763307e26188b227bc175cbad92857717e74e7a1be1215e9095943d
+oid sha256:8c2fafb89d14c402d0168feedd1d7f90498106fee1e65183676fc96e12b98842
 size 73911112

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 1.164143992582467e-06,
-    "train_runtime": 617.3268,
+    "train_loss": 1.293529453505471e-06,
+    "train_runtime": 624.0298,
     "train_samples": 11,
-    "train_samples_per_second": 0.518,
+    "train_samples_per_second": 0.513,
     "train_steps_per_second": 0.032
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 1.164143992582467e-06,
-    "train_runtime": 617.3268,
+    "train_loss": 1.293529453505471e-06,
+    "train_runtime": 624.0298,
     "train_samples": 11,
-    "train_samples_per_second": 0.518,
+    "train_samples_per_second": 0.513,
     "train_steps_per_second": 0.032
 }

trainer_state.json

@@ -10,202 +10,202 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 463.8125,
+      "completion_length": 519.4375,
       "epoch": 0.7272727272727273,
-      "grad_norm": 0.3649120032787323,
-      "kl": 0.0012721834136755206,
+      "grad_norm": 0.573968768119812,
+      "kl": 0.0011488548334455118,
       "learning_rate": 5e-07,
       "loss": 0.0,
-      "reward": 1.022139385342598,
-      "reward_std": 1.1148683652281761,
-      "rewards/concensus_correctness_reward_func": 0.027437500655651093,
-      "rewards/consensus_reward_func": 0.25,
+      "reward": 1.1180308051407337,
+      "reward_std": 1.122335970401764,
+      "rewards/concensus_correctness_reward_func": 0.02931249886751175,
+      "rewards/consensus_reward_func": 0.375,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.47973313741385937,
+      "rewards/question_recreation_reward_func": 0.6167183108627796,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.046875,
-      "rewards/xmlcount_reward_func": 0.21809375658631325,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.09700000286102295,
       "step": 2
     },
     {
-      "completion_length": 535.8571428571429,
+      "completion_length": 439.60714285714283,
       "epoch": 1.3636363636363638,
-      "grad_norm": 3.259235382080078,
-      "kl": 0.0010872288756737752,
+      "grad_norm": 0.8938176035881042,
+      "kl": 0.0015368413545989565,
       "learning_rate": 4.864543104251586e-07,
       "loss": 0.0,
-      "reward": 0.9391504994460514,
-      "reward_std": 1.2883880138397217,
-      "rewards/concensus_correctness_reward_func": 0.0,
+      "reward": 1.0507219859531947,
+      "reward_std": 0.8666104206017086,
+      "rewards/concensus_correctness_reward_func": 0.040214285254478455,
       "rewards/consensus_reward_func": 0.14285714285714285,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.6467219037669045,
+      "rewards/question_recreation_reward_func": 0.5666862768786294,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.1495714315346309,
+      "rewards/strict_format_reward_func": 0.03571428571428571,
+      "rewards/xmlcount_reward_func": 0.2652500037636076,
       "step": 4
     },
     {
-      "completion_length": 443.5357142857143,
+      "completion_length": 479.2142857142857,
       "epoch": 2.0,
-      "grad_norm": 0.3289619982242584,
-      "kl": 0.0013656587190260844,
+      "grad_norm": 0.3144683539867401,
+      "kl": 0.001359426316672138,
       "learning_rate": 4.472851273490984e-07,
       "loss": 0.0,
-      "reward": 0.9847379135234016,
-      "reward_std": 1.0057098737784795,
-      "rewards/concensus_correctness_reward_func": 0.06271428721291679,
+      "reward": 1.1467279302222388,
+      "reward_std": 0.9164375151906695,
+      "rewards/concensus_correctness_reward_func": 0.0,
       "rewards/consensus_reward_func": 0.21428571428571427,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.6736307484763009,
+      "rewards/question_recreation_reward_func": 0.5260136340345655,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.017857142857142856,
-      "rewards/xmlcount_reward_func": 0.016250001532690867,
+      "rewards/xmlcount_reward_func": 0.38857144117355347,
       "step": 6
     },
     {
-      "completion_length": 506.09375,
+      "completion_length": 586.5,
       "epoch": 2.7272727272727275,
-      "grad_norm": 0.38292196393013,
-      "kl": 0.0011815363905043341,
+      "grad_norm": 0.3388758897781372,
+      "kl": 0.0010796373535413295,
       "learning_rate": 3.867370395306068e-07,
       "loss": 0.0,
-      "reward": 1.6399211697280407,
-      "reward_std": 1.5490521118044853,
-      "rewards/concensus_correctness_reward_func": 0.08562500029802322,
-      "rewards/consensus_reward_func": 0.5625,
+      "reward": 1.3970180973410606,
+      "reward_std": 1.1238007210195065,
+      "rewards/concensus_correctness_reward_func": 0.06949999555945396,
+      "rewards/consensus_reward_func": 0.375,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.6976711601018906,
+      "rewards/question_recreation_reward_func": 0.5888618398457766,
       "rewards/soft_format_reward_func": 0.0,
       "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.27850000746548176,
+      "rewards/xmlcount_reward_func": 0.34803124889731407,
       "step": 8
     },
     {
-      "completion_length": 456.5,
+      "completion_length": 512.8214285714286,
       "epoch": 3.3636363636363638,
-      "grad_norm": 0.5194206237792969,
-      "kl": 0.0012395976394015764,
+      "grad_norm": 0.44189029932022095,
+      "kl": 0.0013629934484405176,
       "learning_rate": 3.1137137178519977e-07,
       "loss": 0.0,
-      "reward": 1.257384717464447,
-      "reward_std": 1.1879289150238037,
-      "rewards/concensus_correctness_reward_func": 0.016214285578046526,
+      "reward": 1.2535826787352562,
+      "reward_std": 1.040838543857847,
+      "rewards/concensus_correctness_reward_func": 0.09642857100282397,
       "rewards/consensus_reward_func": 0.2857142857142857,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.595099042568888,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5115469821861812,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.28892856410571505,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.34203568526676725,
       "step": 10
     },
     {
-      "completion_length": 510.14285714285717,
+      "completion_length": 402.14285714285717,
       "epoch": 4.0,
-      "grad_norm": 0.28684717416763306,
-      "kl": 0.0012059293471143714,
+      "grad_norm": 0.4739115536212921,
+      "kl": 0.0015622548775614373,
       "learning_rate": 2.2935516363191693e-07,
       "loss": 0.0,
-      "reward": 0.8977083606379372,
-      "reward_std": 1.0021978744438715,
-      "rewards/concensus_correctness_reward_func": 0.031357143606458394,
-      "rewards/consensus_reward_func": 0.2857142857142857,
+      "reward": 1.1068306054387773,
+      "reward_std": 0.8945644923618862,
+      "rewards/concensus_correctness_reward_func": 0.01242857107094356,
+      "rewards/consensus_reward_func": 0.42857142857142855,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.5461012103727886,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.5422949242804732,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": -0.03689286006348474,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.10567857431513923,
       "step": 12
     },
     {
-      "completion_length": 564.1875,
+      "completion_length": 460.03125,
       "epoch": 4.7272727272727275,
-      "grad_norm": 0.3074406087398529,
-      "kl": 0.001165143468824681,
+      "grad_norm": 0.4402799904346466,
+      "kl": 0.0016441376355942339,
       "learning_rate": 1.4957614383675767e-07,
       "loss": 0.0,
-      "reward": 1.0764202997088432,
-      "reward_std": 0.9078357070684433,
-      "rewards/concensus_correctness_reward_func": 0.0078125,
+      "reward": 1.3537148535251617,
+      "reward_std": 0.98868827521801,
+      "rewards/concensus_correctness_reward_func": 0.04662499949336052,
       "rewards/consensus_reward_func": 0.375,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.6252952925860882,
+      "rewards/question_recreation_reward_func": 0.6408398598432541,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.015625,
-      "rewards/xmlcount_reward_func": 0.05268751084804535,
+      "rewards/strict_format_reward_func": 0.0,
+      "rewards/xmlcount_reward_func": 0.29125000163912773,
       "step": 14
     },
     {
-      "completion_length": 425.57142857142856,
+      "completion_length": 369.42857142857144,
       "epoch": 5.363636363636363,
-      "grad_norm": 0.48346689343452454,
-      "kl": 0.001424723836992468,
+      "grad_norm": 0.535619854927063,
+      "kl": 0.0013823732484265097,
       "learning_rate": 8.067960709356478e-08,
       "loss": 0.0,
-      "reward": 0.9693987795284816,
-      "reward_std": 0.8530563030924115,
-      "rewards/concensus_correctness_reward_func": 0.0,
-      "rewards/consensus_reward_func": 0.14285714285714285,
+      "reward": 1.1202760764530726,
+      "reward_std": 1.2240942546299525,
+      "rewards/concensus_correctness_reward_func": 0.033499998705727715,
+      "rewards/consensus_reward_func": 0.42857142857142855,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.5103273359792573,
+      "rewards/question_recreation_reward_func": 0.5740617790392467,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.3162143017564501,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.06628571184618133,
       "step": 16
     },
     {
-      "completion_length": 504.2142857142857,
+      "completion_length": 409.2142857142857,
       "epoch": 6.0,
-      "grad_norm": 0.2852988839149475,
-      "kl": 0.0011552662056471621,
+      "grad_norm": 0.4284095764160156,
+      "kl": 0.0015758839435875416,
       "learning_rate": 3.013156219837776e-08,
       "loss": 0.0,
-      "reward": 1.8628105010305132,
-      "reward_std": 1.0663678986685616,
-      "rewards/concensus_correctness_reward_func": 0.04757142918450492,
-      "rewards/consensus_reward_func": 0.5714285714285714,
+      "reward": 0.919320170368467,
+      "reward_std": 0.9560085790497916,
+      "rewards/concensus_correctness_reward_func": 0.027785714183534895,
+      "rewards/consensus_reward_func": 0.2857142857142857,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.07142857142857142,
-      "rewards/question_recreation_reward_func": 0.731917679309845,
-      "rewards/soft_format_reward_func": 0.017857142857142856,
-      "rewards/strict_format_reward_func": 0.0,
-      "rewards/xmlcount_reward_func": 0.42260715791157316,
+      "rewards/final_correctness_reward_func": 0.0,
+      "rewards/question_recreation_reward_func": 0.4710344544478825,
+      "rewards/soft_format_reward_func": 0.0,
+      "rewards/strict_format_reward_func": 0.017857142857142856,
+      "rewards/xmlcount_reward_func": 0.1169285774230957,
       "step": 18
     },
     {
-      "completion_length": 414.90625,
+      "completion_length": 527.34375,
       "epoch": 6.7272727272727275,
-      "grad_norm": 0.43161511421203613,
-      "kl": 0.0014090457698330283,
+      "grad_norm": 0.44153422117233276,
+      "kl": 0.0012525435740826651,
       "learning_rate": 3.4096741493194193e-09,
       "loss": 0.0,
-      "reward": 1.1701091527938843,
-      "reward_std": 0.7885858789086342,
-      "rewards/concensus_correctness_reward_func": 0.027437500655651093,
-      "rewards/consensus_reward_func": 0.3125,
+      "reward": 0.7976084109395742,
+      "reward_std": 1.2174969762563705,
+      "rewards/concensus_correctness_reward_func": 0.010874999687075615,
+      "rewards/consensus_reward_func": 0.25,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.0,
-      "rewards/question_recreation_reward_func": 0.4807029031217098,
-      "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.046875,
-      "rewards/xmlcount_reward_func": 0.30259374901652336,
+      "rewards/question_recreation_reward_func": 0.49245213717222214,
+      "rewards/soft_format_reward_func": 0.015625,
+      "rewards/strict_format_reward_func": 0.015625,
+      "rewards/xmlcount_reward_func": 0.013031253591179848,
       "step": 20
     },
     {
       "epoch": 6.7272727272727275,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 1.164143992582467e-06,
-      "train_runtime": 617.3268,
-      "train_samples_per_second": 0.518,
+      "train_loss": 1.293529453505471e-06,
+      "train_runtime": 624.0298,
+      "train_samples_per_second": 0.513,
       "train_steps_per_second": 0.032
     }
   ],