checkpoint-24/trainer_state.json

{
  "best_global_step": 24,
  "best_metric": 0.6869648694992065,
  "best_model_checkpoint": "//outputs/task7_microsoft/Phi-3.5-mini-instruct/checkpoint-24",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 24,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.43478260869565216,
      "grad_norm": 0.3567487597465515,
      "learning_rate": 4e-05,
      "loss": 0.7927,
      "step": 5
    },
    {
      "epoch": 0.8695652173913043,
      "grad_norm": 0.20804046094417572,
      "learning_rate": 8e-05,
      "loss": 0.877,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.7741447687149048,
      "eval_runtime": 3.8872,
      "eval_samples_per_second": 5.145,
      "eval_steps_per_second": 0.772,
      "step": 12
    },
    {
      "epoch": 1.2608695652173914,
      "grad_norm": 0.14667491614818573,
      "learning_rate": 0.00012,
      "loss": 0.797,
      "step": 15
    },
    {
      "epoch": 1.6956521739130435,
      "grad_norm": 0.18651264905929565,
      "learning_rate": 0.00016,
      "loss": 0.6654,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.6869648694992065,
      "eval_runtime": 3.8967,
      "eval_samples_per_second": 5.132,
      "eval_steps_per_second": 0.77,
      "step": 24
    }
  ],
  "logging_steps": 5,
  "max_steps": 50,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "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": 1.4018199449038848e+16,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}