trainer_state.json

{
  "best_metric": 10.546875,
  "best_model_checkpoint": "model_bloomz_7b1_tr_v3_it_2/checkpoint-30000",
  "epoch": 1.75,
  "eval_steps": 6000,
  "global_step": 42000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.25,
      "learning_rate": 1.95e-05,
      "loss": 10.4994,
      "step": 6000
    },
    {
      "epoch": 0.25,
      "eval_loss": 10.5703125,
      "eval_runtime": 5733.8305,
      "eval_samples_per_second": 1.046,
      "eval_steps_per_second": 1.046,
      "step": 6000
    },
    {
      "epoch": 0.5,
      "learning_rate": 1.9e-05,
      "loss": 10.5463,
      "step": 12000
    },
    {
      "epoch": 0.5,
      "eval_loss": 10.5703125,
      "eval_runtime": 4294.5164,
      "eval_samples_per_second": 1.397,
      "eval_steps_per_second": 1.397,
      "step": 12000
    },
    {
      "epoch": 0.75,
      "learning_rate": 1.8500000000000002e-05,
      "loss": 10.5565,
      "step": 18000
    },
    {
      "epoch": 0.75,
      "eval_loss": 10.5703125,
      "eval_runtime": 6332.6473,
      "eval_samples_per_second": 0.947,
      "eval_steps_per_second": 0.947,
      "step": 18000
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.8e-05,
      "loss": 10.5465,
      "step": 24000
    },
    {
      "epoch": 1.0,
      "eval_loss": 10.5546875,
      "eval_runtime": 6326.006,
      "eval_samples_per_second": 0.948,
      "eval_steps_per_second": 0.948,
      "step": 24000
    },
    {
      "epoch": 1.25,
      "learning_rate": 1.7500000000000002e-05,
      "loss": 10.559,
      "step": 30000
    },
    {
      "epoch": 1.25,
      "eval_loss": 10.546875,
      "eval_runtime": 6325.6234,
      "eval_samples_per_second": 0.949,
      "eval_steps_per_second": 0.949,
      "step": 30000
    },
    {
      "epoch": 1.5,
      "learning_rate": 1.7e-05,
      "loss": 10.517,
      "step": 36000
    },
    {
      "epoch": 1.5,
      "eval_loss": 10.546875,
      "eval_runtime": 6331.0478,
      "eval_samples_per_second": 0.948,
      "eval_steps_per_second": 0.948,
      "step": 36000
    },
    {
      "epoch": 1.75,
      "learning_rate": 1.65e-05,
      "loss": 10.4977,
      "step": 42000
    },
    {
      "epoch": 1.75,
      "eval_loss": 10.546875,
      "eval_runtime": 6333.619,
      "eval_samples_per_second": 0.947,
      "eval_steps_per_second": 0.947,
      "step": 42000
    }
  ],
  "logging_steps": 6000,
  "max_steps": 240000,
  "num_train_epochs": 10,
  "save_steps": 6000,
  "total_flos": 3.117948352856064e+18,
  "trial_name": null,
  "trial_params": null
}