trainer_state.json

{
  "best_metric": 1.0906689167022705,
  "best_model_checkpoint": "./results/checkpoint-2699",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 10796,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.6147003173828125,
      "learning_rate": 4.672437244961049e-05,
      "loss": 0.327,
      "step": 2699
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7347354138398915,
      "eval_conf_mat": [
        [
          6123,
          717
        ],
        [
          2802,
          3624
        ]
      ],
      "eval_f1": 0.673168013374199,
      "eval_loss": 1.0906689167022705,
      "eval_precision": 0.8348306841741534,
      "eval_recall": 0.5639589169000934,
      "eval_runtime": 50.9369,
      "eval_samples_per_second": 260.44,
      "eval_steps_per_second": 8.147,
      "step": 2699
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.085843563079834,
      "learning_rate": 4.338691727463831e-05,
      "loss": 0.14,
      "step": 5398
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7100105532941354,
      "eval_conf_mat": [
        [
          6405,
          435
        ],
        [
          3412,
          3014
        ]
      ],
      "eval_f1": 0.6104303797468354,
      "eval_loss": 1.5652275085449219,
      "eval_precision": 0.8738764859379531,
      "eval_recall": 0.46903205726735137,
      "eval_runtime": 51.0274,
      "eval_samples_per_second": 259.978,
      "eval_steps_per_second": 8.133,
      "step": 5398
    },
    {
      "epoch": 3.0,
      "grad_norm": 7.21763277053833,
      "learning_rate": 4.004946209966614e-05,
      "loss": 0.0959,
      "step": 8097
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7397105382180009,
      "eval_conf_mat": [
        [
          6322,
          518
        ],
        [
          2935,
          3491
        ]
      ],
      "eval_f1": 0.6690943938667945,
      "eval_loss": 1.292863130569458,
      "eval_precision": 0.8707907208780244,
      "eval_recall": 0.5432617491441021,
      "eval_runtime": 49.9344,
      "eval_samples_per_second": 265.669,
      "eval_steps_per_second": 8.311,
      "step": 8097
    },
    {
      "epoch": 4.0,
      "grad_norm": 18.75651741027832,
      "learning_rate": 3.6712006924693956e-05,
      "loss": 0.0676,
      "step": 10796
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7308156188753203,
      "eval_conf_mat": [
        [
          6358,
          482
        ],
        [
          3089,
          3337
        ]
      ],
      "eval_f1": 0.6514397266959493,
      "eval_loss": 1.5299923419952393,
      "eval_precision": 0.8737889499869076,
      "eval_recall": 0.5192966075319017,
      "eval_runtime": 51.18,
      "eval_samples_per_second": 259.203,
      "eval_steps_per_second": 8.109,
      "step": 10796
    }
  ],
  "logging_steps": 500,
  "max_steps": 40485,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 15,
  "save_steps": 500,
  "total_flos": 4.192333290845046e+16,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}