{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.9724770642201834,
  "eval_steps": 40,
  "global_step": 81,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1834862385321101,
      "grad_norm": 87.98338604183625,
      "learning_rate": 5e-07,
      "logits/chosen": -2.7504944801330566,
      "logits/rejected": -2.7376608848571777,
      "logps/chosen": -366.7567138671875,
      "logps/rejected": -269.1701965332031,
      "loss": 0.684,
      "rewards/accuracies": 0.38749998807907104,
      "rewards/chosen": 0.036848217248916626,
      "rewards/margins": 0.02083454094827175,
      "rewards/rejected": 0.016013674437999725,
      "step": 5
    },
    {
      "epoch": 0.3669724770642202,
      "grad_norm": 56.51853547131025,
      "learning_rate": 1e-06,
      "logits/chosen": -2.680494785308838,
      "logits/rejected": -2.666748285293579,
      "logps/chosen": -349.07464599609375,
      "logps/rejected": -251.41049194335938,
      "loss": 0.5321,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": 1.2032493352890015,
      "rewards/margins": 0.6572977900505066,
      "rewards/rejected": 0.5459514260292053,
      "step": 10
    },
    {
      "epoch": 0.5504587155963303,
      "grad_norm": 50.68732493997359,
      "learning_rate": 9.878131657762535e-07,
      "logits/chosen": -2.505305290222168,
      "logits/rejected": -2.4723613262176514,
      "logps/chosen": -344.2197265625,
      "logps/rejected": -244.26724243164062,
      "loss": 0.4596,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": 3.3816330432891846,
      "rewards/margins": 1.7416467666625977,
      "rewards/rejected": 1.639986276626587,
      "step": 15
    },
    {
      "epoch": 0.7339449541284404,
      "grad_norm": 48.68084600724118,
      "learning_rate": 9.518467388186019e-07,
      "logits/chosen": -2.3610739707946777,
      "logits/rejected": -2.312293529510498,
      "logps/chosen": -348.7062072753906,
      "logps/rejected": -251.2678985595703,
      "loss": 0.4779,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": 3.990943431854248,
      "rewards/margins": 2.4283714294433594,
      "rewards/rejected": 1.5625723600387573,
      "step": 20
    },
    {
      "epoch": 0.9174311926605505,
      "grad_norm": 43.576923617976306,
      "learning_rate": 8.938539866588592e-07,
      "logits/chosen": -2.161365509033203,
      "logits/rejected": -2.169524669647217,
      "logps/chosen": -337.71240234375,
      "logps/rejected": -223.12796020507812,
      "loss": 0.4429,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": 3.8454651832580566,
      "rewards/margins": 2.679399013519287,
      "rewards/rejected": 1.16606605052948,
      "step": 25
    },
    {
      "epoch": 1.1009174311926606,
      "grad_norm": 15.165089558011688,
      "learning_rate": 8.166619015240235e-07,
      "logits/chosen": -2.103114366531372,
      "logits/rejected": -2.0902843475341797,
      "logps/chosen": -333.57958984375,
      "logps/rejected": -255.9213409423828,
      "loss": 0.2402,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": 4.685088157653809,
      "rewards/margins": 3.5494351387023926,
      "rewards/rejected": 1.135652780532837,
      "step": 30
    },
    {
      "epoch": 1.2844036697247707,
      "grad_norm": 16.2176567972852,
      "learning_rate": 7.240333919937892e-07,
      "logits/chosen": -2.1111814975738525,
      "logits/rejected": -2.107719898223877,
      "logps/chosen": -336.36669921875,
      "logps/rejected": -258.0069885253906,
      "loss": 0.0996,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": 4.797186374664307,
      "rewards/margins": 4.8145318031311035,
      "rewards/rejected": -0.01734566129744053,
      "step": 35
    },
    {
      "epoch": 1.4678899082568808,
      "grad_norm": 23.35951964253066,
      "learning_rate": 6.204838512283071e-07,
      "logits/chosen": -2.1685638427734375,
      "logits/rejected": -2.134986400604248,
      "logps/chosen": -321.1785583496094,
      "logps/rejected": -274.7965087890625,
      "loss": 0.1046,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": 4.402096271514893,
      "rewards/margins": 4.899113178253174,
      "rewards/rejected": -0.4970162808895111,
      "step": 40
    },
    {
      "epoch": 1.4678899082568808,
      "eval_logits/chosen": -2.157338857650757,
      "eval_logits/rejected": -2.1405954360961914,
      "eval_logps/chosen": -318.1454772949219,
      "eval_logps/rejected": -250.12286376953125,
      "eval_loss": 0.41041094064712524,
      "eval_rewards/accuracies": 0.8557692170143127,
      "eval_rewards/chosen": 3.198086977005005,
      "eval_rewards/margins": 2.8902511596679688,
      "eval_rewards/rejected": 0.30783578753471375,
      "eval_runtime": 54.4376,
      "eval_samples_per_second": 14.126,
      "eval_steps_per_second": 0.239,
      "step": 40
    },
    {
      "epoch": 1.6513761467889907,
      "grad_norm": 15.026762247253014,
      "learning_rate": 5.110610435765934e-07,
      "logits/chosen": -2.1868062019348145,
      "logits/rejected": -2.1655430793762207,
      "logps/chosen": -328.25323486328125,
      "logps/rejected": -256.1419677734375,
      "loss": 0.1147,
      "rewards/accuracies": 0.9937499761581421,
      "rewards/chosen": 5.091620445251465,
      "rewards/margins": 5.604047775268555,
      "rewards/rejected": -0.5124271512031555,
      "step": 45
    },
    {
      "epoch": 1.834862385321101,
      "grad_norm": 20.207846152861585,
      "learning_rate": 4.010990395072413e-07,
      "logits/chosen": -2.1973319053649902,
      "logits/rejected": -2.182528018951416,
      "logps/chosen": -318.63800048828125,
      "logps/rejected": -282.4230041503906,
      "loss": 0.1138,
      "rewards/accuracies": 0.987500011920929,
      "rewards/chosen": 4.645673751831055,
      "rewards/margins": 5.373934268951416,
      "rewards/rejected": -0.7282606959342957,
      "step": 50
    },
    {
      "epoch": 2.018348623853211,
      "grad_norm": 11.21625290323588,
      "learning_rate": 2.9595819387826747e-07,
      "logits/chosen": -2.2084720134735107,
      "logits/rejected": -2.195996046066284,
      "logps/chosen": -319.896484375,
      "logps/rejected": -269.0523986816406,
      "loss": 0.0766,
      "rewards/accuracies": 0.9937499761581421,
      "rewards/chosen": 5.313421726226807,
      "rewards/margins": 5.936707496643066,
      "rewards/rejected": -0.6232857704162598,
      "step": 55
    },
    {
      "epoch": 2.2018348623853212,
      "grad_norm": 12.136370494834754,
      "learning_rate": 2.0076384291297133e-07,
      "logits/chosen": -2.221501588821411,
      "logits/rejected": -2.2167458534240723,
      "logps/chosen": -319.49652099609375,
      "logps/rejected": -252.19302368164062,
      "loss": 0.032,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 5.2415971755981445,
      "rewards/margins": 5.633230686187744,
      "rewards/rejected": -0.39163410663604736,
      "step": 60
    },
    {
      "epoch": 2.385321100917431,
      "grad_norm": 12.64835328707971,
      "learning_rate": 1.2015645770835764e-07,
      "logits/chosen": -2.235281467437744,
      "logits/rejected": -2.1987690925598145,
      "logps/chosen": -336.44158935546875,
      "logps/rejected": -248.538330078125,
      "loss": 0.0343,
      "rewards/accuracies": 0.981249988079071,
      "rewards/chosen": 5.285120487213135,
      "rewards/margins": 5.766313552856445,
      "rewards/rejected": -0.48119330406188965,
      "step": 65
    },
    {
      "epoch": 2.5688073394495414,
      "grad_norm": 9.209595589190794,
      "learning_rate": 5.806543362721944e-08,
      "logits/chosen": -2.2242534160614014,
      "logits/rejected": -2.2229883670806885,
      "logps/chosen": -326.62701416015625,
      "logps/rejected": -266.9122619628906,
      "loss": 0.0218,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 5.088566780090332,
      "rewards/margins": 5.991884708404541,
      "rewards/rejected": -0.9033180475234985,
      "step": 70
    },
    {
      "epoch": 2.7522935779816513,
      "grad_norm": 6.119217367247741,
      "learning_rate": 1.751754273859507e-08,
      "logits/chosen": -2.2509617805480957,
      "logits/rejected": -2.230938673019409,
      "logps/chosen": -326.35186767578125,
      "logps/rejected": -290.31903076171875,
      "loss": 0.0466,
      "rewards/accuracies": 0.981249988079071,
      "rewards/chosen": 5.807272911071777,
      "rewards/margins": 6.667996883392334,
      "rewards/rejected": -0.8607242703437805,
      "step": 75
    },
    {
      "epoch": 2.9357798165137616,
      "grad_norm": 13.26496056644377,
      "learning_rate": 4.893867400131979e-10,
      "logits/chosen": -2.244044780731201,
      "logits/rejected": -2.239701986312866,
      "logps/chosen": -341.17584228515625,
      "logps/rejected": -274.3210144042969,
      "loss": 0.0449,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 5.888747215270996,
      "rewards/margins": 6.674208641052246,
      "rewards/rejected": -0.7854617238044739,
      "step": 80
    },
    {
      "epoch": 2.9357798165137616,
      "eval_logits/chosen": -2.2088987827301025,
      "eval_logits/rejected": -2.196852445602417,
      "eval_logps/chosen": -315.8323669433594,
      "eval_logps/rejected": -252.4415740966797,
      "eval_loss": 0.4136447310447693,
      "eval_rewards/accuracies": 0.8557692170143127,
      "eval_rewards/chosen": 3.4293932914733887,
      "eval_rewards/margins": 3.3534319400787354,
      "eval_rewards/rejected": 0.07596174627542496,
      "eval_runtime": 54.1579,
      "eval_samples_per_second": 14.199,
      "eval_steps_per_second": 0.24,
      "step": 80
    },
    {
      "epoch": 2.9724770642201834,
      "step": 81,
      "total_flos": 954757539692544.0,
      "train_loss": 0.21779433532077588,
      "train_runtime": 2944.3137,
      "train_samples_per_second": 7.049,
      "train_steps_per_second": 0.028
    }
  ],
  "logging_steps": 5,
  "max_steps": 81,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 40,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 954757539692544.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}