checkpoint-18/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 18,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.17,
      "grad_norm": NaN,
      "learning_rate": 0.0,
      "loss": 8.9144,
      "step": 1
    },
    {
      "epoch": 0.33,
      "grad_norm": NaN,
      "learning_rate": 0.0,
      "loss": 12.2548,
      "step": 2
    },
    {
      "epoch": 0.5,
      "grad_norm": NaN,
      "learning_rate": 0.0,
      "loss": 8.0577,
      "step": 3
    },
    {
      "epoch": 0.67,
      "grad_norm": Infinity,
      "learning_rate": 0.0,
      "loss": 10.0668,
      "step": 4
    },
    {
      "epoch": 0.83,
      "grad_norm": 118.140869140625,
      "learning_rate": 1.5e-05,
      "loss": 11.3165,
      "step": 5
    },
    {
      "epoch": 1.0,
      "grad_norm": 119.7070541381836,
      "learning_rate": 3e-05,
      "loss": 8.0135,
      "step": 6
    },
    {
      "epoch": 1.17,
      "grad_norm": 73.31137084960938,
      "learning_rate": 2.8125e-05,
      "loss": 7.0185,
      "step": 7
    },
    {
      "epoch": 1.33,
      "grad_norm": 48.86985397338867,
      "learning_rate": 2.625e-05,
      "loss": 6.366,
      "step": 8
    },
    {
      "epoch": 1.5,
      "grad_norm": 63.3082389831543,
      "learning_rate": 2.4375e-05,
      "loss": 6.9771,
      "step": 9
    },
    {
      "epoch": 1.67,
      "grad_norm": 60.80664825439453,
      "learning_rate": 2.25e-05,
      "loss": 4.8249,
      "step": 10
    },
    {
      "epoch": 1.83,
      "grad_norm": 48.479217529296875,
      "learning_rate": 2.0625e-05,
      "loss": 4.9635,
      "step": 11
    },
    {
      "epoch": 2.0,
      "grad_norm": 24.786556243896484,
      "learning_rate": 1.8750000000000002e-05,
      "loss": 3.3671,
      "step": 12
    },
    {
      "epoch": 2.17,
      "grad_norm": 16.406631469726562,
      "learning_rate": 1.6875e-05,
      "loss": 2.361,
      "step": 13
    },
    {
      "epoch": 2.33,
      "grad_norm": 15.527493476867676,
      "learning_rate": 1.5e-05,
      "loss": 3.1927,
      "step": 14
    },
    {
      "epoch": 2.5,
      "grad_norm": 28.2950382232666,
      "learning_rate": 1.3125e-05,
      "loss": 4.0086,
      "step": 15
    },
    {
      "epoch": 2.67,
      "grad_norm": 25.018417358398438,
      "learning_rate": 1.125e-05,
      "loss": 3.4131,
      "step": 16
    },
    {
      "epoch": 2.83,
      "grad_norm": 15.129682540893555,
      "learning_rate": 9.375000000000001e-06,
      "loss": 2.6968,
      "step": 17
    },
    {
      "epoch": 3.0,
      "grad_norm": 13.875788688659668,
      "learning_rate": 7.5e-06,
      "loss": 2.714,
      "step": 18
    }
  ],
  "logging_steps": 1,
  "max_steps": 18,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 6902048951894016.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}