trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9852216748768473,
  "eval_steps": 100,
  "global_step": 400,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.12,
      "learning_rate": 2.3109243697478994e-05,
      "loss": 1.0832,
      "step": 50
    },
    {
      "epoch": 0.25,
      "learning_rate": 2.100840336134454e-05,
      "loss": 0.4234,
      "step": 100
    },
    {
      "epoch": 0.25,
      "eval_loss": 0.36679086089134216,
      "eval_runtime": 111.544,
      "eval_samples_per_second": 1.219,
      "eval_steps_per_second": 0.152,
      "step": 100
    },
    {
      "epoch": 0.37,
      "learning_rate": 1.8907563025210083e-05,
      "loss": 0.3757,
      "step": 150
    },
    {
      "epoch": 0.49,
      "learning_rate": 1.6806722689075634e-05,
      "loss": 0.3644,
      "step": 200
    },
    {
      "epoch": 0.49,
      "eval_loss": 0.32610517740249634,
      "eval_runtime": 111.5555,
      "eval_samples_per_second": 1.219,
      "eval_steps_per_second": 0.152,
      "step": 200
    },
    {
      "epoch": 0.62,
      "learning_rate": 1.4705882352941177e-05,
      "loss": 0.3444,
      "step": 250
    },
    {
      "epoch": 0.74,
      "learning_rate": 1.2605042016806723e-05,
      "loss": 0.3289,
      "step": 300
    },
    {
      "epoch": 0.74,
      "eval_loss": 0.3098466098308563,
      "eval_runtime": 111.5402,
      "eval_samples_per_second": 1.219,
      "eval_steps_per_second": 0.152,
      "step": 300
    },
    {
      "epoch": 0.86,
      "learning_rate": 1.050420168067227e-05,
      "loss": 0.3027,
      "step": 350
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.403361344537817e-06,
      "loss": 0.2973,
      "step": 400
    },
    {
      "epoch": 0.99,
      "eval_loss": 0.2937043309211731,
      "eval_runtime": 111.5631,
      "eval_samples_per_second": 1.219,
      "eval_steps_per_second": 0.152,
      "step": 400
    }
  ],
  "logging_steps": 50,
  "max_steps": 600,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 100,
  "total_flos": 4.6557229848576e+17,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}