{
  "best_metric": 0.9547275641025641,
  "best_model_checkpoint": "text/train/checkpoint/pubmedbert-fulltext/ddi_42/checkpoint-7119",
  "epoch": 10.0,
  "eval_steps": 500,
  "global_step": 7910,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9471153846153846,
      "eval_loss": 0.19469894468784332,
      "eval_runtime": 5.2082,
      "eval_samples_per_second": 479.246,
      "eval_steps_per_second": 1.92,
      "step": 791
    },
    {
      "epoch": 1.26,
      "grad_norm": 8.167149543762207,
      "learning_rate": 4.367888748419722e-05,
      "loss": 0.1709,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9527243589743589,
      "eval_loss": 0.2473936378955841,
      "eval_runtime": 4.9302,
      "eval_samples_per_second": 506.267,
      "eval_steps_per_second": 2.028,
      "step": 1582
    },
    {
      "epoch": 2.53,
      "grad_norm": 0.3635082244873047,
      "learning_rate": 3.735777496839444e-05,
      "loss": 0.0734,
      "step": 2000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9475160256410257,
      "eval_loss": 0.24853640794754028,
      "eval_runtime": 4.9236,
      "eval_samples_per_second": 506.945,
      "eval_steps_per_second": 2.031,
      "step": 2373
    },
    {
      "epoch": 3.79,
      "grad_norm": 7.960566520690918,
      "learning_rate": 3.1036662452591655e-05,
      "loss": 0.0475,
      "step": 3000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9499198717948718,
      "eval_loss": 0.26862233877182007,
      "eval_runtime": 4.8986,
      "eval_samples_per_second": 509.536,
      "eval_steps_per_second": 2.041,
      "step": 3164
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9475160256410257,
      "eval_loss": 0.3195907175540924,
      "eval_runtime": 4.8903,
      "eval_samples_per_second": 510.395,
      "eval_steps_per_second": 2.045,
      "step": 3955
    },
    {
      "epoch": 5.06,
      "grad_norm": 0.0017318876925855875,
      "learning_rate": 2.4715549936788876e-05,
      "loss": 0.0284,
      "step": 4000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9527243589743589,
      "eval_loss": 0.3014063239097595,
      "eval_runtime": 4.8924,
      "eval_samples_per_second": 510.174,
      "eval_steps_per_second": 2.044,
      "step": 4746
    },
    {
      "epoch": 6.32,
      "grad_norm": 0.025161907076835632,
      "learning_rate": 1.8394437420986094e-05,
      "loss": 0.0194,
      "step": 5000
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.952323717948718,
      "eval_loss": 0.3125186860561371,
      "eval_runtime": 4.9579,
      "eval_samples_per_second": 503.443,
      "eval_steps_per_second": 2.017,
      "step": 5537
    },
    {
      "epoch": 7.59,
      "grad_norm": 0.21671424806118011,
      "learning_rate": 1.2073324905183313e-05,
      "loss": 0.0133,
      "step": 6000
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9491185897435898,
      "eval_loss": 0.3640977442264557,
      "eval_runtime": 4.9394,
      "eval_samples_per_second": 505.32,
      "eval_steps_per_second": 2.025,
      "step": 6328
    },
    {
      "epoch": 8.85,
      "grad_norm": 0.002933135023340583,
      "learning_rate": 5.752212389380531e-06,
      "loss": 0.0065,
      "step": 7000
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9547275641025641,
      "eval_loss": 0.3299635350704193,
      "eval_runtime": 4.8971,
      "eval_samples_per_second": 509.686,
      "eval_steps_per_second": 2.042,
      "step": 7119
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9543269230769231,
      "eval_loss": 0.3501527011394501,
      "eval_runtime": 4.9496,
      "eval_samples_per_second": 504.282,
      "eval_steps_per_second": 2.02,
      "step": 7910
    },
    {
      "epoch": 10.0,
      "step": 7910,
      "total_flos": 1.663959132942336e+16,
      "train_loss": 0.04588771306458677,
      "train_runtime": 1428.2798,
      "train_samples_per_second": 177.108,
      "train_steps_per_second": 5.538
    }
  ],
  "logging_steps": 1000,
  "max_steps": 7910,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 1.663959132942336e+16,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}