{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.9927360774818403,
  "eval_steps": 500,
  "global_step": 618,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.12,
      "learning_rate": 2.0161290322580642e-07,
      "logps/chosen": -155.47128295898438,
      "logps/rejected": -104.8045883178711,
      "loss": 0.7536,
      "losses/dpo": 0.7361882925033569,
      "losses/sft": 1.3545048236846924,
      "losses/total": 0.7361882925033569,
      "ref_logps/chosen": -155.36932373046875,
      "ref_logps/rejected": -104.93973541259766,
      "rewards/accuracies": 0.48450005054473877,
      "rewards/chosen": -0.010196244344115257,
      "rewards/margins": -0.023710083216428757,
      "rewards/rejected": 0.0135138388723135,
      "step": 25
    },
    {
      "epoch": 0.24,
      "learning_rate": 4.0322580645161285e-07,
      "logps/chosen": -153.7573699951172,
      "logps/rejected": -107.0270004272461,
      "loss": 0.7401,
      "losses/dpo": 0.7450882792472839,
      "losses/sft": 1.3266704082489014,
      "losses/total": 0.7450882792472839,
      "ref_logps/chosen": -153.8381805419922,
      "ref_logps/rejected": -107.06134033203125,
      "rewards/accuracies": 0.502500057220459,
      "rewards/chosen": 0.00808356050401926,
      "rewards/margins": 0.004649868700653315,
      "rewards/rejected": 0.003433691570535302,
      "step": 50
    },
    {
      "epoch": 0.36,
      "learning_rate": 4.883093525179856e-07,
      "logps/chosen": -151.94009399414062,
      "logps/rejected": -105.40670013427734,
      "loss": 0.7213,
      "losses/dpo": 0.7041549682617188,
      "losses/sft": 1.3607875108718872,
      "losses/total": 0.7041549682617188,
      "ref_logps/chosen": -152.60272216796875,
      "ref_logps/rejected": -105.59368896484375,
      "rewards/accuracies": 0.5264999866485596,
      "rewards/chosen": 0.06626255810260773,
      "rewards/margins": 0.04756266996264458,
      "rewards/rejected": 0.01869989186525345,
      "step": 75
    },
    {
      "epoch": 0.48,
      "learning_rate": 4.6582733812949637e-07,
      "logps/chosen": -149.70230102539062,
      "logps/rejected": -106.14666748046875,
      "loss": 0.6625,
      "losses/dpo": 0.6611780524253845,
      "losses/sft": 1.341475009918213,
      "losses/total": 0.6611780524253845,
      "ref_logps/chosen": -152.02659606933594,
      "ref_logps/rejected": -106.69593811035156,
      "rewards/accuracies": 0.621999979019165,
      "rewards/chosen": 0.2324293702840805,
      "rewards/margins": 0.17750366032123566,
      "rewards/rejected": 0.05492572858929634,
      "step": 100
    },
    {
      "epoch": 0.61,
      "learning_rate": 4.4334532374100717e-07,
      "logps/chosen": -148.40090942382812,
      "logps/rejected": -104.86791229248047,
      "loss": 0.597,
      "losses/dpo": 0.5830589532852173,
      "losses/sft": 1.2935805320739746,
      "losses/total": 0.5830589532852173,
      "ref_logps/chosen": -152.3332977294922,
      "ref_logps/rejected": -105.52149200439453,
      "rewards/accuracies": 0.7135000824928284,
      "rewards/chosen": 0.39323729276657104,
      "rewards/margins": 0.32787904143333435,
      "rewards/rejected": 0.0653582364320755,
      "step": 125
    },
    {
      "epoch": 0.73,
      "learning_rate": 4.2086330935251797e-07,
      "logps/chosen": -147.95791625976562,
      "logps/rejected": -106.17552947998047,
      "loss": 0.5515,
      "losses/dpo": 0.5297109484672546,
      "losses/sft": 1.289250135421753,
      "losses/total": 0.5297109484672546,
      "ref_logps/chosen": -153.35595703125,
      "ref_logps/rejected": -106.98403930664062,
      "rewards/accuracies": 0.7540000677108765,
      "rewards/chosen": 0.5398061871528625,
      "rewards/margins": 0.45895519852638245,
      "rewards/rejected": 0.0808510109782219,
      "step": 150
    },
    {
      "epoch": 0.85,
      "learning_rate": 3.9838129496402877e-07,
      "logps/chosen": -147.28538513183594,
      "logps/rejected": -106.73645782470703,
      "loss": 0.5171,
      "losses/dpo": 0.526803731918335,
      "losses/sft": 1.2763028144836426,
      "losses/total": 0.526803731918335,
      "ref_logps/chosen": -153.7596893310547,
      "ref_logps/rejected": -107.54974365234375,
      "rewards/accuracies": 0.7815000414848328,
      "rewards/chosen": 0.6474303603172302,
      "rewards/margins": 0.5661011934280396,
      "rewards/rejected": 0.0813291072845459,
      "step": 175
    },
    {
      "epoch": 0.97,
      "learning_rate": 3.7589928057553957e-07,
      "logps/chosen": -146.67481994628906,
      "logps/rejected": -104.6893539428711,
      "loss": 0.4544,
      "losses/dpo": 0.4492318332195282,
      "losses/sft": 1.284212350845337,
      "losses/total": 0.4492318332195282,
      "ref_logps/chosen": -154.76519775390625,
      "ref_logps/rejected": -105.09065246582031,
      "rewards/accuracies": 0.825499951839447,
      "rewards/chosen": 0.809037446975708,
      "rewards/margins": 0.7689078450202942,
      "rewards/rejected": 0.040129706263542175,
      "step": 200
    },
    {
      "epoch": 1.09,
      "learning_rate": 3.5341726618705037e-07,
      "logps/chosen": -144.06991577148438,
      "logps/rejected": -105.81298828125,
      "loss": 0.4189,
      "losses/dpo": 0.42898857593536377,
      "losses/sft": 1.2516932487487793,
      "losses/total": 0.42898857593536377,
      "ref_logps/chosen": -153.25570678710938,
      "ref_logps/rejected": -106.06501007080078,
      "rewards/accuracies": 0.8469999432563782,
      "rewards/chosen": 0.9185801148414612,
      "rewards/margins": 0.8933786749839783,
      "rewards/rejected": 0.02520136535167694,
      "step": 225
    },
    {
      "epoch": 1.21,
      "learning_rate": 3.309352517985611e-07,
      "logps/chosen": -144.94163513183594,
      "logps/rejected": -106.42683410644531,
      "loss": 0.3982,
      "losses/dpo": 0.41044384241104126,
      "losses/sft": 1.2750244140625,
      "losses/total": 0.41044384241104126,
      "ref_logps/chosen": -154.86021423339844,
      "ref_logps/rejected": -106.34297180175781,
      "rewards/accuracies": 0.8530000448226929,
      "rewards/chosen": 0.9918593764305115,
      "rewards/margins": 1.0002468824386597,
      "rewards/rejected": -0.008387637324631214,
      "step": 250
    },
    {
      "epoch": 1.33,
      "learning_rate": 3.0845323741007197e-07,
      "logps/chosen": -144.10977172851562,
      "logps/rejected": -105.52392578125,
      "loss": 0.3652,
      "losses/dpo": 0.3518593907356262,
      "losses/sft": 1.2418614625930786,
      "losses/total": 0.3518593907356262,
      "ref_logps/chosen": -154.4581298828125,
      "ref_logps/rejected": -104.62450408935547,
      "rewards/accuracies": 0.8764999508857727,
      "rewards/chosen": 1.0348376035690308,
      "rewards/margins": 1.1247813701629639,
      "rewards/rejected": -0.08994373679161072,
      "step": 275
    },
    {
      "epoch": 1.45,
      "learning_rate": 2.859712230215827e-07,
      "logps/chosen": -142.1906280517578,
      "logps/rejected": -108.99441528320312,
      "loss": 0.3624,
      "losses/dpo": 0.35955867171287537,
      "losses/sft": 1.2526122331619263,
      "losses/total": 0.35955867171287537,
      "ref_logps/chosen": -152.7955322265625,
      "ref_logps/rejected": -107.89449310302734,
      "rewards/accuracies": 0.8740000128746033,
      "rewards/chosen": 1.0604920387268066,
      "rewards/margins": 1.170485496520996,
      "rewards/rejected": -0.10999350249767303,
      "step": 300
    },
    {
      "epoch": 1.57,
      "learning_rate": 2.634892086330935e-07,
      "logps/chosen": -141.9879913330078,
      "logps/rejected": -107.29058837890625,
      "loss": 0.349,
      "losses/dpo": 0.3294835686683655,
      "losses/sft": 1.2578469514846802,
      "losses/total": 0.3294835686683655,
      "ref_logps/chosen": -152.93646240234375,
      "ref_logps/rejected": -105.73262786865234,
      "rewards/accuracies": 0.8819999694824219,
      "rewards/chosen": 1.0948452949523926,
      "rewards/margins": 1.2506405115127563,
      "rewards/rejected": -0.15579518675804138,
      "step": 325
    },
    {
      "epoch": 1.69,
      "learning_rate": 2.410071942446043e-07,
      "logps/chosen": -142.55641174316406,
      "logps/rejected": -108.43463134765625,
      "loss": 0.3091,
      "losses/dpo": 0.3135191798210144,
      "losses/sft": 1.2662179470062256,
      "losses/total": 0.3135191798210144,
      "ref_logps/chosen": -154.17462158203125,
      "ref_logps/rejected": -105.654052734375,
      "rewards/accuracies": 0.8995000720024109,
      "rewards/chosen": 1.1618221998214722,
      "rewards/margins": 1.4398791790008545,
      "rewards/rejected": -0.2780568599700928,
      "step": 350
    },
    {
      "epoch": 1.82,
      "learning_rate": 2.185251798561151e-07,
      "logps/chosen": -141.23895263671875,
      "logps/rejected": -108.13822937011719,
      "loss": 0.3071,
      "losses/dpo": 0.299526184797287,
      "losses/sft": 1.2248637676239014,
      "losses/total": 0.299526184797287,
      "ref_logps/chosen": -153.57838439941406,
      "ref_logps/rejected": -105.29540252685547,
      "rewards/accuracies": 0.8889999389648438,
      "rewards/chosen": 1.2339427471160889,
      "rewards/margins": 1.5182260274887085,
      "rewards/rejected": -0.28428351879119873,
      "step": 375
    },
    {
      "epoch": 1.94,
      "learning_rate": 1.960431654676259e-07,
      "logps/chosen": -143.84800720214844,
      "logps/rejected": -109.23783111572266,
      "loss": 0.3079,
      "losses/dpo": 0.31919562816619873,
      "losses/sft": 1.2433806657791138,
      "losses/total": 0.31919562816619873,
      "ref_logps/chosen": -155.57464599609375,
      "ref_logps/rejected": -105.85549926757812,
      "rewards/accuracies": 0.8924999237060547,
      "rewards/chosen": 1.1726653575897217,
      "rewards/margins": 1.510898470878601,
      "rewards/rejected": -0.33823302388191223,
      "step": 400
    },
    {
      "epoch": 2.06,
      "learning_rate": 1.735611510791367e-07,
      "logps/chosen": -142.04263305664062,
      "logps/rejected": -110.5551528930664,
      "loss": 0.2964,
      "losses/dpo": 0.29985514283180237,
      "losses/sft": 1.256807565689087,
      "losses/total": 0.29985514283180237,
      "ref_logps/chosen": -153.96051025390625,
      "ref_logps/rejected": -106.5252914428711,
      "rewards/accuracies": 0.8864999413490295,
      "rewards/chosen": 1.1917892694473267,
      "rewards/margins": 1.5947766304016113,
      "rewards/rejected": -0.4029873013496399,
      "step": 425
    },
    {
      "epoch": 2.18,
      "learning_rate": 1.5107913669064747e-07,
      "logps/chosen": -141.11489868164062,
      "logps/rejected": -110.66223907470703,
      "loss": 0.2921,
      "losses/dpo": 0.29900702834129333,
      "losses/sft": 1.223684549331665,
      "losses/total": 0.29900702834129333,
      "ref_logps/chosen": -153.32489013671875,
      "ref_logps/rejected": -106.36825561523438,
      "rewards/accuracies": 0.9029999375343323,
      "rewards/chosen": 1.2209986448287964,
      "rewards/margins": 1.6503956317901611,
      "rewards/rejected": -0.42939692735671997,
      "step": 450
    },
    {
      "epoch": 2.3,
      "learning_rate": 1.2859712230215827e-07,
      "logps/chosen": -140.06715393066406,
      "logps/rejected": -111.3785629272461,
      "loss": 0.2771,
      "losses/dpo": 0.2982315719127655,
      "losses/sft": 1.2262225151062012,
      "losses/total": 0.2982315719127655,
      "ref_logps/chosen": -152.20010375976562,
      "ref_logps/rejected": -106.2849349975586,
      "rewards/accuracies": 0.8995000720024109,
      "rewards/chosen": 1.2132951021194458,
      "rewards/margins": 1.722659707069397,
      "rewards/rejected": -0.509364664554596,
      "step": 475
    },
    {
      "epoch": 2.42,
      "learning_rate": 1.0611510791366907e-07,
      "logps/chosen": -142.41409301757812,
      "logps/rejected": -112.02964782714844,
      "loss": 0.2692,
      "losses/dpo": 0.26506972312927246,
      "losses/sft": 1.2326300144195557,
      "losses/total": 0.26506972312927246,
      "ref_logps/chosen": -154.93585205078125,
      "ref_logps/rejected": -106.74388885498047,
      "rewards/accuracies": 0.9050000905990601,
      "rewards/chosen": 1.252176284790039,
      "rewards/margins": 1.7807520627975464,
      "rewards/rejected": -0.5285759568214417,
      "step": 500
    },
    {
      "epoch": 2.54,
      "learning_rate": 8.363309352517984e-08,
      "logps/chosen": -141.35763549804688,
      "logps/rejected": -111.21255493164062,
      "loss": 0.2639,
      "losses/dpo": 0.2448505461215973,
      "losses/sft": 1.2256782054901123,
      "losses/total": 0.2448505461215973,
      "ref_logps/chosen": -153.86410522460938,
      "ref_logps/rejected": -105.44302368164062,
      "rewards/accuracies": 0.9070001244544983,
      "rewards/chosen": 1.250647783279419,
      "rewards/margins": 1.8276008367538452,
      "rewards/rejected": -0.576953113079071,
      "step": 525
    },
    {
      "epoch": 2.66,
      "learning_rate": 6.115107913669064e-08,
      "logps/chosen": -140.52743530273438,
      "logps/rejected": -111.19217681884766,
      "loss": 0.253,
      "losses/dpo": 0.2362019419670105,
      "losses/sft": 1.219558835029602,
      "losses/total": 0.2362019419670105,
      "ref_logps/chosen": -153.1510009765625,
      "ref_logps/rejected": -104.75523376464844,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.2623597383499146,
      "rewards/margins": 1.9060544967651367,
      "rewards/rejected": -0.6436949372291565,
      "step": 550
    },
    {
      "epoch": 2.78,
      "learning_rate": 3.866906474820144e-08,
      "logps/chosen": -142.2875213623047,
      "logps/rejected": -113.81513977050781,
      "loss": 0.2529,
      "losses/dpo": 0.24827653169631958,
      "losses/sft": 1.2383555173873901,
      "losses/total": 0.24827653169631958,
      "ref_logps/chosen": -154.96365356445312,
      "ref_logps/rejected": -107.25000762939453,
      "rewards/accuracies": 0.9150000214576721,
      "rewards/chosen": 1.2676140069961548,
      "rewards/margins": 1.9241262674331665,
      "rewards/rejected": -0.656512439250946,
      "step": 575
    },
    {
      "epoch": 2.91,
      "learning_rate": 1.618705035971223e-08,
      "logps/chosen": -140.9930877685547,
      "logps/rejected": -113.36444091796875,
      "loss": 0.2651,
      "losses/dpo": 0.27665671706199646,
      "losses/sft": 1.2188746929168701,
      "losses/total": 0.27665671706199646,
      "ref_logps/chosen": -153.5058135986328,
      "ref_logps/rejected": -107.16682434082031,
      "rewards/accuracies": 0.9030001163482666,
      "rewards/chosen": 1.251274585723877,
      "rewards/margins": 1.871036410331726,
      "rewards/rejected": -0.6197616457939148,
      "step": 600
    },
    {
      "epoch": 2.99,
      "step": 618,
      "total_flos": 0.0,
      "train_loss": 0.41181609237078326,
      "train_runtime": 7630.1245,
      "train_samples_per_second": 6.494,
      "train_steps_per_second": 0.081
    }
  ],
  "logging_steps": 25,
  "max_steps": 618,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {},
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}