{
  "best_metric": 0.14963118731975555,
  "best_model_checkpoint": "subtaskA/0/checkpoint-8621",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 25863,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.06,
      "learning_rate": 4.903336813208058e-06,
      "loss": 0.3323,
      "step": 500
    },
    {
      "epoch": 0.12,
      "learning_rate": 4.806673626416116e-06,
      "loss": 0.1967,
      "step": 1000
    },
    {
      "epoch": 0.17,
      "learning_rate": 4.710010439624174e-06,
      "loss": 0.1751,
      "step": 1500
    },
    {
      "epoch": 0.23,
      "learning_rate": 4.613347252832231e-06,
      "loss": 0.1455,
      "step": 2000
    },
    {
      "epoch": 0.29,
      "learning_rate": 4.516684066040289e-06,
      "loss": 0.1402,
      "step": 2500
    },
    {
      "epoch": 0.35,
      "learning_rate": 4.420020879248348e-06,
      "loss": 0.1204,
      "step": 3000
    },
    {
      "epoch": 0.41,
      "learning_rate": 4.323357692456405e-06,
      "loss": 0.121,
      "step": 3500
    },
    {
      "epoch": 0.46,
      "learning_rate": 4.226694505664463e-06,
      "loss": 0.1137,
      "step": 4000
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.130031318872521e-06,
      "loss": 0.1053,
      "step": 4500
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.033368132080579e-06,
      "loss": 0.1108,
      "step": 5000
    },
    {
      "epoch": 0.64,
      "learning_rate": 3.936704945288636e-06,
      "loss": 0.1065,
      "step": 5500
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.840041758496694e-06,
      "loss": 0.1126,
      "step": 6000
    },
    {
      "epoch": 0.75,
      "learning_rate": 3.743378571704752e-06,
      "loss": 0.0865,
      "step": 6500
    },
    {
      "epoch": 0.81,
      "learning_rate": 3.64671538491281e-06,
      "loss": 0.092,
      "step": 7000
    },
    {
      "epoch": 0.87,
      "learning_rate": 3.550052198120868e-06,
      "loss": 0.0961,
      "step": 7500
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.453389011328926e-06,
      "loss": 0.0967,
      "step": 8000
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.356725824536984e-06,
      "loss": 0.0922,
      "step": 8500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9706820554460039,
      "eval_loss": 0.14963118731975555,
      "eval_runtime": 893.6565,
      "eval_samples_per_second": 38.588,
      "eval_steps_per_second": 2.413,
      "step": 8621
    },
    {
      "epoch": 1.04,
      "learning_rate": 3.260062637745041e-06,
      "loss": 0.067,
      "step": 9000
    },
    {
      "epoch": 1.1,
      "learning_rate": 3.1633994509530992e-06,
      "loss": 0.0621,
      "step": 9500
    },
    {
      "epoch": 1.16,
      "learning_rate": 3.066736264161157e-06,
      "loss": 0.0657,
      "step": 10000
    },
    {
      "epoch": 1.22,
      "learning_rate": 2.970073077369215e-06,
      "loss": 0.0641,
      "step": 10500
    },
    {
      "epoch": 1.28,
      "learning_rate": 2.873409890577273e-06,
      "loss": 0.0525,
      "step": 11000
    },
    {
      "epoch": 1.33,
      "learning_rate": 2.7767467037853306e-06,
      "loss": 0.0603,
      "step": 11500
    },
    {
      "epoch": 1.39,
      "learning_rate": 2.6800835169933887e-06,
      "loss": 0.0581,
      "step": 12000
    },
    {
      "epoch": 1.45,
      "learning_rate": 2.5834203302014464e-06,
      "loss": 0.0536,
      "step": 12500
    },
    {
      "epoch": 1.51,
      "learning_rate": 2.486757143409504e-06,
      "loss": 0.0583,
      "step": 13000
    },
    {
      "epoch": 1.57,
      "learning_rate": 2.390093956617562e-06,
      "loss": 0.0564,
      "step": 13500
    },
    {
      "epoch": 1.62,
      "learning_rate": 2.2934307698256196e-06,
      "loss": 0.0633,
      "step": 14000
    },
    {
      "epoch": 1.68,
      "learning_rate": 2.1967675830336777e-06,
      "loss": 0.0544,
      "step": 14500
    },
    {
      "epoch": 1.74,
      "learning_rate": 2.1001043962417355e-06,
      "loss": 0.056,
      "step": 15000
    },
    {
      "epoch": 1.8,
      "learning_rate": 2.003441209449793e-06,
      "loss": 0.0526,
      "step": 15500
    },
    {
      "epoch": 1.86,
      "learning_rate": 1.9067780226578511e-06,
      "loss": 0.0503,
      "step": 16000
    },
    {
      "epoch": 1.91,
      "learning_rate": 1.810114835865909e-06,
      "loss": 0.0519,
      "step": 16500
    },
    {
      "epoch": 1.97,
      "learning_rate": 1.713451649073967e-06,
      "loss": 0.0473,
      "step": 17000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9578645168773924,
      "eval_loss": 0.2317158281803131,
      "eval_runtime": 893.2092,
      "eval_samples_per_second": 38.607,
      "eval_steps_per_second": 2.414,
      "step": 17242
    },
    {
      "epoch": 2.03,
      "learning_rate": 1.6167884622820245e-06,
      "loss": 0.0379,
      "step": 17500
    },
    {
      "epoch": 2.09,
      "learning_rate": 1.5201252754900824e-06,
      "loss": 0.0207,
      "step": 18000
    },
    {
      "epoch": 2.15,
      "learning_rate": 1.4234620886981404e-06,
      "loss": 0.0322,
      "step": 18500
    },
    {
      "epoch": 2.2,
      "learning_rate": 1.326798901906198e-06,
      "loss": 0.0306,
      "step": 19000
    },
    {
      "epoch": 2.26,
      "learning_rate": 1.230135715114256e-06,
      "loss": 0.0273,
      "step": 19500
    },
    {
      "epoch": 2.32,
      "learning_rate": 1.133472528322314e-06,
      "loss": 0.0305,
      "step": 20000
    },
    {
      "epoch": 2.38,
      "learning_rate": 1.0368093415303717e-06,
      "loss": 0.0303,
      "step": 20500
    },
    {
      "epoch": 2.44,
      "learning_rate": 9.401461547384296e-07,
      "loss": 0.0301,
      "step": 21000
    },
    {
      "epoch": 2.49,
      "learning_rate": 8.434829679464873e-07,
      "loss": 0.0203,
      "step": 21500
    },
    {
      "epoch": 2.55,
      "learning_rate": 7.468197811545453e-07,
      "loss": 0.0292,
      "step": 22000
    },
    {
      "epoch": 2.61,
      "learning_rate": 6.50156594362603e-07,
      "loss": 0.0286,
      "step": 22500
    },
    {
      "epoch": 2.67,
      "learning_rate": 5.534934075706608e-07,
      "loss": 0.0318,
      "step": 23000
    },
    {
      "epoch": 2.73,
      "learning_rate": 4.568302207787187e-07,
      "loss": 0.0262,
      "step": 23500
    },
    {
      "epoch": 2.78,
      "learning_rate": 3.6016703398677656e-07,
      "loss": 0.0231,
      "step": 24000
    },
    {
      "epoch": 2.84,
      "learning_rate": 2.6350384719483434e-07,
      "loss": 0.0229,
      "step": 24500
    },
    {
      "epoch": 2.9,
      "learning_rate": 1.668406604028922e-07,
      "loss": 0.0237,
      "step": 25000
    },
    {
      "epoch": 2.96,
      "learning_rate": 7.017747361095002e-08,
      "loss": 0.0202,
      "step": 25500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9638093028650968,
      "eval_loss": 0.2504064738750458,
      "eval_runtime": 892.9889,
      "eval_samples_per_second": 38.616,
      "eval_steps_per_second": 2.414,
      "step": 25863
    }
  ],
  "logging_steps": 500,
  "max_steps": 25863,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 3.855619345186869e+17,
  "trial_name": null,
  "trial_params": null
}