trainer_state.json

{
  "best_metric": 0.11141396313905716,
  "best_model_checkpoint": "./vit-spam/checkpoint-358",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 358,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.13966480446927373,
      "grad_norm": 1.035046935081482,
      "learning_rate": 4.301675977653631e-05,
      "loss": 0.2956,
      "step": 50
    },
    {
      "epoch": 0.27932960893854747,
      "grad_norm": 2.2606287002563477,
      "learning_rate": 3.603351955307263e-05,
      "loss": 0.1171,
      "step": 100
    },
    {
      "epoch": 0.41899441340782123,
      "grad_norm": 1.3705707788467407,
      "learning_rate": 2.9050279329608944e-05,
      "loss": 0.1116,
      "step": 150
    },
    {
      "epoch": 0.5586592178770949,
      "grad_norm": 1.3012491464614868,
      "learning_rate": 2.206703910614525e-05,
      "loss": 0.0839,
      "step": 200
    },
    {
      "epoch": 0.6983240223463687,
      "grad_norm": 0.7185125350952148,
      "learning_rate": 1.5083798882681566e-05,
      "loss": 0.069,
      "step": 250
    },
    {
      "epoch": 0.8379888268156425,
      "grad_norm": 0.9136179685592651,
      "learning_rate": 8.100558659217877e-06,
      "loss": 0.0885,
      "step": 300
    },
    {
      "epoch": 0.9776536312849162,
      "grad_norm": 0.8141963481903076,
      "learning_rate": 1.11731843575419e-06,
      "loss": 0.0785,
      "step": 350
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.11141396313905716,
      "eval_roc_auc": 0.9529222871380703,
      "eval_runtime": 162.1567,
      "eval_samples_per_second": 34.553,
      "eval_steps_per_second": 0.543,
      "step": 358
    },
    {
      "epoch": 1.0,
      "step": 358,
      "total_flos": 1.772241802485719e+18,
      "train_loss": 0.1194506121081347,
      "train_runtime": 815.7328,
      "train_samples_per_second": 28.036,
      "train_steps_per_second": 0.439
    }
  ],
  "logging_steps": 50,
  "max_steps": 358,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 1.772241802485719e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}