checkpoint-500/trainer_state.json

{
  "best_metric": 0.4999743402004242,
  "best_model_checkpoint": "mikhail-panzo/zlm-fil_b64_le5_s8000/checkpoint-500",
  "epoch": 22.22222222222222,
  "eval_steps": 500,
  "global_step": 500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 4.803515434265137,
      "learning_rate": 2.5000000000000004e-07,
      "loss": 0.8968,
      "step": 50
    },
    {
      "epoch": 4.444444444444445,
      "grad_norm": 2.6444506645202637,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.8539,
      "step": 100
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 1.9820537567138672,
      "learning_rate": 7.5e-07,
      "loss": 0.7856,
      "step": 150
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 1.2031773328781128,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.7306,
      "step": 200
    },
    {
      "epoch": 11.11111111111111,
      "grad_norm": 1.728398084640503,
      "learning_rate": 1.25e-06,
      "loss": 0.6821,
      "step": 250
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 1.232366681098938,
      "learning_rate": 1.5e-06,
      "loss": 0.6348,
      "step": 300
    },
    {
      "epoch": 15.555555555555555,
      "grad_norm": 1.4039260149002075,
      "learning_rate": 1.75e-06,
      "loss": 0.5962,
      "step": 350
    },
    {
      "epoch": 17.77777777777778,
      "grad_norm": 1.1848849058151245,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.5758,
      "step": 400
    },
    {
      "epoch": 20.0,
      "grad_norm": 0.9365352392196655,
      "learning_rate": 2.25e-06,
      "loss": 0.5636,
      "step": 450
    },
    {
      "epoch": 22.22222222222222,
      "grad_norm": 0.9893051981925964,
      "learning_rate": 2.5e-06,
      "loss": 0.5529,
      "step": 500
    },
    {
      "epoch": 22.22222222222222,
      "eval_loss": 0.4999743402004242,
      "eval_runtime": 6.8701,
      "eval_samples_per_second": 23.144,
      "eval_steps_per_second": 2.911,
      "step": 500
    }
  ],
  "logging_steps": 50,
  "max_steps": 8000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 364,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 7259852606272992.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}