trainer_state.json

{
  "best_metric": 1.6593416929244995,
  "best_model_checkpoint": "output/viktor-tsoi/checkpoint-82",
  "epoch": 2.0,
  "global_step": 82,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.12,
      "learning_rate": 0.0001322266681456037,
      "loss": 2.2855,
      "step": 5
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00011802778132101399,
      "loss": 2.0776,
      "step": 10
    },
    {
      "epoch": 0.37,
      "learning_rate": 9.666210850995393e-05,
      "loss": 1.9417,
      "step": 15
    },
    {
      "epoch": 0.49,
      "learning_rate": 7.122756753113643e-05,
      "loss": 1.7633,
      "step": 20
    },
    {
      "epoch": 0.61,
      "learning_rate": 4.541204214117672e-05,
      "loss": 1.9391,
      "step": 25
    },
    {
      "epoch": 0.73,
      "learning_rate": 2.2958656968642224e-05,
      "loss": 1.7733,
      "step": 30
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.123042792471594e-06,
      "loss": 1.864,
      "step": 35
    },
    {
      "epoch": 0.98,
      "learning_rate": 2.0128603879540573e-07,
      "loss": 1.8436,
      "step": 40
    },
    {
      "epoch": 1.0,
      "eval_loss": 1.742375135421753,
      "eval_runtime": 2.5783,
      "eval_samples_per_second": 20.944,
      "eval_steps_per_second": 2.715,
      "step": 41
    },
    {
      "epoch": 1.1,
      "learning_rate": 3.197007505031765e-06,
      "loss": 1.858,
      "step": 45
    },
    {
      "epoch": 1.22,
      "learning_rate": 1.5675842264214697e-05,
      "loss": 1.8206,
      "step": 50
    },
    {
      "epoch": 1.34,
      "learning_rate": 3.58284204500588e-05,
      "loss": 1.7518,
      "step": 55
    },
    {
      "epoch": 1.46,
      "learning_rate": 6.0732717017669706e-05,
      "loss": 1.744,
      "step": 60
    },
    {
      "epoch": 1.59,
      "learning_rate": 8.677773105069102e-05,
      "loss": 1.7542,
      "step": 65
    },
    {
      "epoch": 1.71,
      "learning_rate": 0.00011018706319231134,
      "loss": 1.7612,
      "step": 70
    },
    {
      "epoch": 1.83,
      "learning_rate": 0.00012756647503932202,
      "loss": 1.7551,
      "step": 75
    },
    {
      "epoch": 1.95,
      "learning_rate": 0.0001363960370713319,
      "loss": 1.7343,
      "step": 80
    },
    {
      "epoch": 2.0,
      "eval_loss": 1.6593416929244995,
      "eval_runtime": 2.6581,
      "eval_samples_per_second": 21.444,
      "eval_steps_per_second": 3.01,
      "step": 82
    }
  ],
  "max_steps": 82,
  "num_train_epochs": 2,
  "total_flos": 84397326336000.0,
  "trial_name": null,
  "trial_params": null
}