trainer_state.json

{
  "best_metric": 2.128573179244995,
  "best_model_checkpoint": "output/nirvana/checkpoint-72",
  "epoch": 2.0,
  "global_step": 72,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.14,
      "learning_rate": 0.00013040646433810595,
      "loss": 2.8877,
      "step": 5
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00011137140040750913,
      "loss": 2.7784,
      "step": 10
    },
    {
      "epoch": 0.43,
      "learning_rate": 8.386493606940316e-05,
      "loss": 2.2959,
      "step": 15
    },
    {
      "epoch": 0.57,
      "learning_rate": 5.3335063930596836e-05,
      "loss": 2.5076,
      "step": 20
    },
    {
      "epoch": 0.71,
      "learning_rate": 2.5828599592490882e-05,
      "loss": 2.4917,
      "step": 25
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.793535661894054e-06,
      "loss": 2.4772,
      "step": 30
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0,
      "loss": 2.6382,
      "step": 35
    },
    {
      "epoch": 1.0,
      "eval_loss": 2.5190556049346924,
      "eval_runtime": 2.4884,
      "eval_samples_per_second": 22.504,
      "eval_steps_per_second": 2.813,
      "step": 35
    },
    {
      "epoch": 1.0,
      "eval_loss": 2.3264973163604736,
      "eval_runtime": 1.1647,
      "eval_samples_per_second": 45.506,
      "eval_steps_per_second": 6.01,
      "step": 36
    },
    {
      "epoch": 1.11,
      "learning_rate": 4.137086214086682e-06,
      "loss": 2.6774,
      "step": 40
    },
    {
      "epoch": 1.25,
      "learning_rate": 2.009247481060283e-05,
      "loss": 2.4577,
      "step": 45
    },
    {
      "epoch": 1.39,
      "learning_rate": 4.513741816785908e-05,
      "loss": 2.2823,
      "step": 50
    },
    {
      "epoch": 1.53,
      "learning_rate": 7.457888395248933e-05,
      "loss": 2.298,
      "step": 55
    },
    {
      "epoch": 1.67,
      "learning_rate": 0.00010290000000000001,
      "loss": 2.4088,
      "step": 60
    },
    {
      "epoch": 1.81,
      "learning_rate": 0.00012479383023822482,
      "loss": 2.2944,
      "step": 65
    },
    {
      "epoch": 1.94,
      "learning_rate": 0.00013615781185663748,
      "loss": 2.0032,
      "step": 70
    },
    {
      "epoch": 2.0,
      "eval_loss": 2.128573179244995,
      "eval_runtime": 1.1938,
      "eval_samples_per_second": 44.396,
      "eval_steps_per_second": 5.864,
      "step": 72
    }
  ],
  "max_steps": 72,
  "num_train_epochs": 2,
  "total_flos": 73161768960000.0,
  "trial_name": null,
  "trial_params": null
}