trainer_state.json

{
  "best_global_step": 400,
  "best_metric": 0.9678384954963574,
  "best_model_checkpoint": "BTX24/hiera-finetuned-busi/checkpoint-400",
  "epoch": 48.0,
  "eval_steps": 100,
  "global_step": 480,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 5.0,
      "grad_norm": 10.227621078491211,
      "learning_rate": 1.9999735576321776e-05,
      "loss": 0.6534,
      "step": 50
    },
    {
      "epoch": 10.0,
      "grad_norm": 8.809008598327637,
      "learning_rate": 1.9346189253489888e-05,
      "loss": 0.2276,
      "step": 100
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9615384615384616,
      "eval_f1": 0.9610639894395369,
      "eval_loss": 0.16655339300632477,
      "eval_precision": 0.964046822742475,
      "eval_recall": 0.9615384615384616,
      "eval_runtime": 2.0085,
      "eval_samples_per_second": 77.671,
      "eval_steps_per_second": 9.958,
      "step": 100
    },
    {
      "epoch": 15.0,
      "grad_norm": 5.131218910217285,
      "learning_rate": 1.7470250712409963e-05,
      "loss": 0.1848,
      "step": 150
    },
    {
      "epoch": 20.0,
      "grad_norm": 27.55030059814453,
      "learning_rate": 1.4617486132350343e-05,
      "loss": 0.1447,
      "step": 200
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.9358974358974359,
      "eval_f1": 0.9351834928758006,
      "eval_loss": 0.18257668614387512,
      "eval_precision": 0.9425747863247864,
      "eval_recall": 0.9358974358974359,
      "eval_runtime": 1.8774,
      "eval_samples_per_second": 83.092,
      "eval_steps_per_second": 10.653,
      "step": 200
    },
    {
      "epoch": 25.0,
      "grad_norm": 4.577565670013428,
      "learning_rate": 1.1160929141252303e-05,
      "loss": 0.1148,
      "step": 250
    },
    {
      "epoch": 30.0,
      "grad_norm": 7.3700270652771,
      "learning_rate": 7.623141076738271e-06,
      "loss": 0.0882,
      "step": 300
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.9615384615384616,
      "eval_f1": 0.9611905417400026,
      "eval_loss": 0.16116316616535187,
      "eval_precision": 0.9612915234951416,
      "eval_recall": 0.9615384615384616,
      "eval_runtime": 1.8625,
      "eval_samples_per_second": 83.757,
      "eval_steps_per_second": 10.738,
      "step": 300
    },
    {
      "epoch": 35.0,
      "grad_norm": 1.2752153873443604,
      "learning_rate": 4.323957196315714e-06,
      "loss": 0.0728,
      "step": 350
    },
    {
      "epoch": 40.0,
      "grad_norm": 5.241044044494629,
      "learning_rate": 1.7669848724331984e-06,
      "loss": 0.0606,
      "step": 400
    },
    {
      "epoch": 40.0,
      "eval_accuracy": 0.967948717948718,
      "eval_f1": 0.9678384954963574,
      "eval_loss": 0.1380443274974823,
      "eval_precision": 0.9678410301752478,
      "eval_recall": 0.967948717948718,
      "eval_runtime": 1.8643,
      "eval_samples_per_second": 83.679,
      "eval_steps_per_second": 10.728,
      "step": 400
    },
    {
      "epoch": 45.0,
      "grad_norm": 13.990592956542969,
      "learning_rate": 2.865793018673857e-07,
      "loss": 0.0628,
      "step": 450
    },
    {
      "epoch": 48.0,
      "step": 480,
      "total_flos": 1.3730970154696704e+18,
      "train_loss": 0.17117963035901387,
      "train_runtime": 645.3789,
      "train_samples_per_second": 46.41,
      "train_steps_per_second": 0.744
    }
  ],
  "logging_steps": 50,
  "max_steps": 480,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 48,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.3730970154696704e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}