{
  "best_metric": 0.4623475670814514,
  "best_model_checkpoint": "./Vit-GPT2-COCO2017Flickr-80k-08/checkpoint-3500",
  "epoch": 2.1,
  "eval_steps": 500,
  "global_step": 10500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1,
      "grad_norm": 0.9981909990310669,
      "learning_rate": 4.8333333333333334e-05,
      "loss": 0.3691,
      "step": 500
    },
    {
      "epoch": 0.1,
      "eval_gen_len": 11.77575,
      "eval_loss": 0.4729623794555664,
      "eval_rouge1": 39.8086,
      "eval_rouge2": 14.7674,
      "eval_rougeL": 36.1546,
      "eval_rougeLsum": 36.1739,
      "eval_runtime": 459.5392,
      "eval_samples_per_second": 8.704,
      "eval_steps_per_second": 2.176,
      "step": 500
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.8220647573471069,
      "learning_rate": 4.666666666666667e-05,
      "loss": 0.3706,
      "step": 1000
    },
    {
      "epoch": 0.2,
      "eval_gen_len": 11.59775,
      "eval_loss": 0.47387805581092834,
      "eval_rouge1": 39.8972,
      "eval_rouge2": 14.9064,
      "eval_rougeL": 36.1193,
      "eval_rougeLsum": 36.138,
      "eval_runtime": 446.0956,
      "eval_samples_per_second": 8.967,
      "eval_steps_per_second": 2.242,
      "step": 1000
    },
    {
      "epoch": 0.3,
      "grad_norm": 1.052058219909668,
      "learning_rate": 4.5e-05,
      "loss": 0.3709,
      "step": 1500
    },
    {
      "epoch": 0.3,
      "eval_gen_len": 11.71025,
      "eval_loss": 0.47592678666114807,
      "eval_rouge1": 39.9874,
      "eval_rouge2": 14.8528,
      "eval_rougeL": 36.3155,
      "eval_rougeLsum": 36.3317,
      "eval_runtime": 466.5502,
      "eval_samples_per_second": 8.574,
      "eval_steps_per_second": 2.143,
      "step": 1500
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.8642559051513672,
      "learning_rate": 4.3333333333333334e-05,
      "loss": 0.3721,
      "step": 2000
    },
    {
      "epoch": 0.4,
      "eval_gen_len": 12.175,
      "eval_loss": 0.4677760601043701,
      "eval_rouge1": 39.7192,
      "eval_rouge2": 14.5844,
      "eval_rougeL": 35.8447,
      "eval_rougeLsum": 35.8728,
      "eval_runtime": 455.6895,
      "eval_samples_per_second": 8.778,
      "eval_steps_per_second": 2.194,
      "step": 2000
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.8769400715827942,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.3655,
      "step": 2500
    },
    {
      "epoch": 0.5,
      "eval_gen_len": 11.90025,
      "eval_loss": 0.4683643877506256,
      "eval_rouge1": 40.3132,
      "eval_rouge2": 15.1157,
      "eval_rougeL": 36.5749,
      "eval_rougeLsum": 36.5823,
      "eval_runtime": 449.7304,
      "eval_samples_per_second": 8.894,
      "eval_steps_per_second": 2.224,
      "step": 2500
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.8613843321800232,
      "learning_rate": 4e-05,
      "loss": 0.3623,
      "step": 3000
    },
    {
      "epoch": 0.6,
      "eval_gen_len": 12.025,
      "eval_loss": 0.4671708345413208,
      "eval_rouge1": 40.1643,
      "eval_rouge2": 14.978,
      "eval_rougeL": 36.3002,
      "eval_rougeLsum": 36.3232,
      "eval_runtime": 451.1814,
      "eval_samples_per_second": 8.866,
      "eval_steps_per_second": 2.216,
      "step": 3000
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.8106393814086914,
      "learning_rate": 3.8333333333333334e-05,
      "loss": 0.3676,
      "step": 3500
    },
    {
      "epoch": 0.7,
      "eval_gen_len": 11.815,
      "eval_loss": 0.4623475670814514,
      "eval_rouge1": 40.5036,
      "eval_rouge2": 15.3751,
      "eval_rougeL": 36.8369,
      "eval_rougeLsum": 36.867,
      "eval_runtime": 458.4461,
      "eval_samples_per_second": 8.725,
      "eval_steps_per_second": 2.181,
      "step": 3500
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.914828360080719,
      "learning_rate": 3.6666666666666666e-05,
      "loss": 0.3613,
      "step": 4000
    },
    {
      "epoch": 0.8,
      "eval_gen_len": 12.054,
      "eval_loss": 0.46470996737480164,
      "eval_rouge1": 40.4078,
      "eval_rouge2": 15.3105,
      "eval_rougeL": 36.65,
      "eval_rougeLsum": 36.6732,
      "eval_runtime": 457.1563,
      "eval_samples_per_second": 8.75,
      "eval_steps_per_second": 2.187,
      "step": 4000
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.8004090189933777,
      "learning_rate": 3.5e-05,
      "loss": 0.3539,
      "step": 4500
    },
    {
      "epoch": 0.9,
      "eval_gen_len": 11.904,
      "eval_loss": 0.46335914731025696,
      "eval_rouge1": 40.3794,
      "eval_rouge2": 15.233,
      "eval_rougeL": 36.7155,
      "eval_rougeLsum": 36.7435,
      "eval_runtime": 445.5568,
      "eval_samples_per_second": 8.978,
      "eval_steps_per_second": 2.244,
      "step": 4500
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.876920223236084,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.3481,
      "step": 5000
    },
    {
      "epoch": 1.0,
      "eval_gen_len": 11.738,
      "eval_loss": 0.46441808342933655,
      "eval_rouge1": 40.037,
      "eval_rouge2": 14.8477,
      "eval_rougeL": 36.3648,
      "eval_rougeLsum": 36.3903,
      "eval_runtime": 442.7978,
      "eval_samples_per_second": 9.033,
      "eval_steps_per_second": 2.258,
      "step": 5000
    },
    {
      "epoch": 1.1,
      "grad_norm": 0.9966852068901062,
      "learning_rate": 3.1666666666666666e-05,
      "loss": 0.2889,
      "step": 5500
    },
    {
      "epoch": 1.1,
      "eval_gen_len": 11.55,
      "eval_loss": 0.4897365868091583,
      "eval_rouge1": 40.1394,
      "eval_rouge2": 14.7595,
      "eval_rougeL": 36.4428,
      "eval_rougeLsum": 36.4696,
      "eval_runtime": 431.6652,
      "eval_samples_per_second": 9.266,
      "eval_steps_per_second": 2.317,
      "step": 5500
    },
    {
      "epoch": 1.2,
      "grad_norm": 0.857021152973175,
      "learning_rate": 3e-05,
      "loss": 0.2908,
      "step": 6000
    },
    {
      "epoch": 1.2,
      "eval_gen_len": 11.98225,
      "eval_loss": 0.48649004101753235,
      "eval_rouge1": 40.0479,
      "eval_rouge2": 14.8181,
      "eval_rougeL": 36.316,
      "eval_rougeLsum": 36.3519,
      "eval_runtime": 438.3532,
      "eval_samples_per_second": 9.125,
      "eval_steps_per_second": 2.281,
      "step": 6000
    },
    {
      "epoch": 1.3,
      "grad_norm": 0.8857397437095642,
      "learning_rate": 2.8333333333333335e-05,
      "loss": 0.2882,
      "step": 6500
    },
    {
      "epoch": 1.3,
      "eval_gen_len": 11.7945,
      "eval_loss": 0.48633432388305664,
      "eval_rouge1": 40.5912,
      "eval_rouge2": 15.3128,
      "eval_rougeL": 36.7638,
      "eval_rougeLsum": 36.7755,
      "eval_runtime": 432.2098,
      "eval_samples_per_second": 9.255,
      "eval_steps_per_second": 2.314,
      "step": 6500
    },
    {
      "epoch": 1.4,
      "grad_norm": 0.792189359664917,
      "learning_rate": 2.6666666666666667e-05,
      "loss": 0.2901,
      "step": 7000
    },
    {
      "epoch": 1.4,
      "eval_gen_len": 11.87,
      "eval_loss": 0.48679661750793457,
      "eval_rouge1": 40.3138,
      "eval_rouge2": 14.9695,
      "eval_rougeL": 36.5032,
      "eval_rougeLsum": 36.5211,
      "eval_runtime": 436.9846,
      "eval_samples_per_second": 9.154,
      "eval_steps_per_second": 2.288,
      "step": 7000
    },
    {
      "epoch": 1.5,
      "grad_norm": 0.8871281147003174,
      "learning_rate": 2.5e-05,
      "loss": 0.2857,
      "step": 7500
    },
    {
      "epoch": 1.5,
      "eval_gen_len": 11.776,
      "eval_loss": 0.4834408462047577,
      "eval_rouge1": 40.2242,
      "eval_rouge2": 14.9881,
      "eval_rougeL": 36.5381,
      "eval_rougeLsum": 36.5607,
      "eval_runtime": 452.1322,
      "eval_samples_per_second": 8.847,
      "eval_steps_per_second": 2.212,
      "step": 7500
    },
    {
      "epoch": 1.6,
      "grad_norm": 1.0286399126052856,
      "learning_rate": 2.3333333333333336e-05,
      "loss": 0.279,
      "step": 8000
    },
    {
      "epoch": 1.6,
      "eval_gen_len": 12.01325,
      "eval_loss": 0.4999020993709564,
      "eval_rouge1": 40.2751,
      "eval_rouge2": 15.0173,
      "eval_rougeL": 36.4172,
      "eval_rougeLsum": 36.4257,
      "eval_runtime": 435.9247,
      "eval_samples_per_second": 9.176,
      "eval_steps_per_second": 2.294,
      "step": 8000
    },
    {
      "epoch": 1.7,
      "grad_norm": 0.9107884168624878,
      "learning_rate": 2.1666666666666667e-05,
      "loss": 0.281,
      "step": 8500
    },
    {
      "epoch": 1.7,
      "eval_gen_len": 11.7685,
      "eval_loss": 0.49508431553840637,
      "eval_rouge1": 40.1172,
      "eval_rouge2": 14.8119,
      "eval_rougeL": 36.2966,
      "eval_rougeLsum": 36.296,
      "eval_runtime": 425.7116,
      "eval_samples_per_second": 9.396,
      "eval_steps_per_second": 2.349,
      "step": 8500
    },
    {
      "epoch": 1.8,
      "grad_norm": 0.76063472032547,
      "learning_rate": 2e-05,
      "loss": 0.2831,
      "step": 9000
    },
    {
      "epoch": 1.8,
      "eval_gen_len": 12.22925,
      "eval_loss": 0.49789097905158997,
      "eval_rouge1": 39.9913,
      "eval_rouge2": 14.7427,
      "eval_rougeL": 36.1539,
      "eval_rougeLsum": 36.1517,
      "eval_runtime": 449.7047,
      "eval_samples_per_second": 8.895,
      "eval_steps_per_second": 2.224,
      "step": 9000
    },
    {
      "epoch": 1.9,
      "grad_norm": 0.9408590197563171,
      "learning_rate": 1.8333333333333333e-05,
      "loss": 0.2799,
      "step": 9500
    },
    {
      "epoch": 1.9,
      "eval_gen_len": 11.87175,
      "eval_loss": 0.49114319682121277,
      "eval_rouge1": 40.5123,
      "eval_rouge2": 15.09,
      "eval_rougeL": 36.7528,
      "eval_rougeLsum": 36.7622,
      "eval_runtime": 440.8811,
      "eval_samples_per_second": 9.073,
      "eval_steps_per_second": 2.268,
      "step": 9500
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.9572837948799133,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 0.2778,
      "step": 10000
    },
    {
      "epoch": 2.0,
      "eval_gen_len": 12.02625,
      "eval_loss": 0.4928734600543976,
      "eval_rouge1": 40.5005,
      "eval_rouge2": 15.1027,
      "eval_rougeL": 36.6202,
      "eval_rougeLsum": 36.6327,
      "eval_runtime": 443.0336,
      "eval_samples_per_second": 9.029,
      "eval_steps_per_second": 2.257,
      "step": 10000
    },
    {
      "epoch": 2.1,
      "grad_norm": 0.9376727938652039,
      "learning_rate": 1.5e-05,
      "loss": 0.2318,
      "step": 10500
    },
    {
      "epoch": 2.1,
      "eval_gen_len": 12.133,
      "eval_loss": 0.5237044095993042,
      "eval_rouge1": 40.1565,
      "eval_rouge2": 14.8022,
      "eval_rougeL": 36.1946,
      "eval_rougeLsum": 36.2074,
      "eval_runtime": 434.3272,
      "eval_samples_per_second": 9.21,
      "eval_steps_per_second": 2.302,
      "step": 10500
    }
  ],
  "logging_steps": 500,
  "max_steps": 15000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "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": 3.0317933708181504e+19,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}