{
  "best_metric": 0.0006635845638811588,
  "best_model_checkpoint": "./megumin_results/checkpoint-2500",
  "epoch": 4.805382027871215,
  "eval_steps": 500,
  "global_step": 2500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.19221528111484862,
      "grad_norm": 4.030124664306641,
      "learning_rate": 3.96e-06,
      "loss": 1.2166,
      "step": 100
    },
    {
      "epoch": 0.38443056222969724,
      "grad_norm": 0.825600266456604,
      "learning_rate": 7.960000000000002e-06,
      "loss": 0.2925,
      "step": 200
    },
    {
      "epoch": 0.5766458433445459,
      "grad_norm": 0.7094722986221313,
      "learning_rate": 1.1920000000000001e-05,
      "loss": 0.0769,
      "step": 300
    },
    {
      "epoch": 0.7688611244593945,
      "grad_norm": 0.23390699923038483,
      "learning_rate": 1.5920000000000003e-05,
      "loss": 0.0445,
      "step": 400
    },
    {
      "epoch": 0.9610764055742431,
      "grad_norm": 0.8664096593856812,
      "learning_rate": 1.9920000000000002e-05,
      "loss": 0.0304,
      "step": 500
    },
    {
      "epoch": 0.9610764055742431,
      "eval_loss": 0.0065709324553608894,
      "eval_runtime": 116.7786,
      "eval_samples_per_second": 15.842,
      "eval_steps_per_second": 1.987,
      "step": 500
    },
    {
      "epoch": 1.1532916866890919,
      "grad_norm": 0.970707356929779,
      "learning_rate": 1.9892723329629885e-05,
      "loss": 0.0215,
      "step": 600
    },
    {
      "epoch": 1.3455069678039404,
      "grad_norm": 0.19242490828037262,
      "learning_rate": 1.9564504327386318e-05,
      "loss": 0.0154,
      "step": 700
    },
    {
      "epoch": 1.5377222489187892,
      "grad_norm": 0.1774681955575943,
      "learning_rate": 1.9022630103336575e-05,
      "loss": 0.0161,
      "step": 800
    },
    {
      "epoch": 1.7299375300336377,
      "grad_norm": 0.5783206224441528,
      "learning_rate": 1.8279205232225727e-05,
      "loss": 0.0127,
      "step": 900
    },
    {
      "epoch": 1.9221528111484862,
      "grad_norm": 0.2338998168706894,
      "learning_rate": 1.7350836597201767e-05,
      "loss": 0.0118,
      "step": 1000
    },
    {
      "epoch": 1.9221528111484862,
      "eval_loss": 0.001669618533924222,
      "eval_runtime": 116.8653,
      "eval_samples_per_second": 15.83,
      "eval_steps_per_second": 1.985,
      "step": 1000
    },
    {
      "epoch": 2.114368092263335,
      "grad_norm": 0.14343367516994476,
      "learning_rate": 1.625826241948815e-05,
      "loss": 0.0091,
      "step": 1100
    },
    {
      "epoch": 2.3065833733781838,
      "grad_norm": 0.2039719969034195,
      "learning_rate": 1.502588900079051e-05,
      "loss": 0.0094,
      "step": 1200
    },
    {
      "epoch": 2.498798654493032,
      "grad_norm": 0.15303683280944824,
      "learning_rate": 1.3681245526846782e-05,
      "loss": 0.0082,
      "step": 1300
    },
    {
      "epoch": 2.691013935607881,
      "grad_norm": 0.07792511582374573,
      "learning_rate": 1.2254369110908413e-05,
      "loss": 0.0083,
      "step": 1400
    },
    {
      "epoch": 2.8832292167227296,
      "grad_norm": 0.07674787193536758,
      "learning_rate": 1.0777133814265167e-05,
      "loss": 0.0073,
      "step": 1500
    },
    {
      "epoch": 2.8832292167227296,
      "eval_loss": 0.0009687193087302148,
      "eval_runtime": 116.8145,
      "eval_samples_per_second": 15.837,
      "eval_steps_per_second": 1.986,
      "step": 1500
    },
    {
      "epoch": 3.075444497837578,
      "grad_norm": 0.25639790296554565,
      "learning_rate": 9.282538632386208e-06,
      "loss": 0.0063,
      "step": 1600
    },
    {
      "epoch": 3.2676597789524267,
      "grad_norm": 0.21751342713832855,
      "learning_rate": 7.80397035189052e-06,
      "loss": 0.0056,
      "step": 1700
    },
    {
      "epoch": 3.4598750600672754,
      "grad_norm": 0.14014215767383575,
      "learning_rate": 6.3744577449038415e-06,
      "loss": 0.0049,
      "step": 1800
    },
    {
      "epoch": 3.6520903411821237,
      "grad_norm": 1.104333519935608,
      "learning_rate": 5.025933760867782e-06,
      "loss": 0.0059,
      "step": 1900
    },
    {
      "epoch": 3.8443056222969725,
      "grad_norm": 0.16992008686065674,
      "learning_rate": 3.7885221972168974e-06,
      "loss": 0.005,
      "step": 2000
    },
    {
      "epoch": 3.8443056222969725,
      "eval_loss": 0.0007113968022167683,
      "eval_runtime": 116.804,
      "eval_samples_per_second": 15.838,
      "eval_steps_per_second": 1.986,
      "step": 2000
    },
    {
      "epoch": 4.036520903411821,
      "grad_norm": 0.9393255114555359,
      "learning_rate": 2.689864783522098e-06,
      "loss": 0.0054,
      "step": 2100
    },
    {
      "epoch": 4.22873618452667,
      "grad_norm": 0.24965894222259521,
      "learning_rate": 1.7545037109285946e-06,
      "loss": 0.005,
      "step": 2200
    },
    {
      "epoch": 4.420951465641519,
      "grad_norm": 0.2450593262910843,
      "learning_rate": 1.0033334001604833e-06,
      "loss": 0.0059,
      "step": 2300
    },
    {
      "epoch": 4.6131667467563675,
      "grad_norm": 0.5003035068511963,
      "learning_rate": 4.5313375468875155e-07,
      "loss": 0.005,
      "step": 2400
    },
    {
      "epoch": 4.805382027871215,
      "grad_norm": 0.019921617582440376,
      "learning_rate": 1.1619532541569333e-07,
      "loss": 0.0042,
      "step": 2500
    },
    {
      "epoch": 4.805382027871215,
      "eval_loss": 0.0006635845638811588,
      "eval_runtime": 116.6762,
      "eval_samples_per_second": 15.856,
      "eval_steps_per_second": 1.988,
      "step": 2500
    }
  ],
  "logging_steps": 100,
  "max_steps": 2600,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 5.822138548224e+16,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}