trainer_state.json

{
  "best_global_step": 1400,
  "best_metric": 0.025447050109505653,
  "best_model_checkpoint": "/kaggle/working/llava-fast-ft-final/checkpoint-1400",
  "epoch": 0.7425742574257426,
  "eval_steps": 200,
  "global_step": 1500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.09900990099009901,
      "grad_norm": 0.07653062790632248,
      "learning_rate": 0.00017413333333333336,
      "loss": 0.8722,
      "step": 200
    },
    {
      "epoch": 0.09900990099009901,
      "eval_loss": 0.03575189411640167,
      "eval_runtime": 1566.5235,
      "eval_samples_per_second": 1.289,
      "eval_steps_per_second": 0.162,
      "step": 200
    },
    {
      "epoch": 0.19801980198019803,
      "grad_norm": 0.059320081025362015,
      "learning_rate": 0.00014746666666666666,
      "loss": 0.0329,
      "step": 400
    },
    {
      "epoch": 0.19801980198019803,
      "eval_loss": 0.03128737956285477,
      "eval_runtime": 1558.9418,
      "eval_samples_per_second": 1.295,
      "eval_steps_per_second": 0.162,
      "step": 400
    },
    {
      "epoch": 0.297029702970297,
      "grad_norm": 0.06010338291525841,
      "learning_rate": 0.0001208,
      "loss": 0.0303,
      "step": 600
    },
    {
      "epoch": 0.297029702970297,
      "eval_loss": 0.029264075681567192,
      "eval_runtime": 1554.6483,
      "eval_samples_per_second": 1.299,
      "eval_steps_per_second": 0.163,
      "step": 600
    },
    {
      "epoch": 0.39603960396039606,
      "grad_norm": 0.05657634884119034,
      "learning_rate": 9.413333333333334e-05,
      "loss": 0.0285,
      "step": 800
    },
    {
      "epoch": 0.39603960396039606,
      "eval_loss": 0.027970999479293823,
      "eval_runtime": 1556.7576,
      "eval_samples_per_second": 1.297,
      "eval_steps_per_second": 0.163,
      "step": 800
    },
    {
      "epoch": 0.49504950495049505,
      "grad_norm": 0.054556068032979965,
      "learning_rate": 6.746666666666666e-05,
      "loss": 0.0276,
      "step": 1000
    },
    {
      "epoch": 0.49504950495049505,
      "eval_loss": 0.02712135948240757,
      "eval_runtime": 1559.148,
      "eval_samples_per_second": 1.295,
      "eval_steps_per_second": 0.162,
      "step": 1000
    },
    {
      "epoch": 0.594059405940594,
      "grad_norm": 0.05568067356944084,
      "learning_rate": 4.08e-05,
      "loss": 0.0263,
      "step": 1200
    },
    {
      "epoch": 0.594059405940594,
      "eval_loss": 0.026194194331765175,
      "eval_runtime": 1555.2028,
      "eval_samples_per_second": 1.298,
      "eval_steps_per_second": 0.163,
      "step": 1200
    },
    {
      "epoch": 0.693069306930693,
      "grad_norm": 0.0598338283598423,
      "learning_rate": 1.4133333333333334e-05,
      "loss": 0.0253,
      "step": 1400
    },
    {
      "epoch": 0.693069306930693,
      "eval_loss": 0.025447050109505653,
      "eval_runtime": 1556.1287,
      "eval_samples_per_second": 1.297,
      "eval_steps_per_second": 0.163,
      "step": 1400
    }
  ],
  "logging_steps": 200,
  "max_steps": 1500,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 200,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 3.6993137678155776e+17,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}