trainer_state.json

{
  "best_metric": 0.8906666666666667,
  "best_model_checkpoint": "./resnet_finetuned_models_dataset/CIFAR100/50_from_100/microsoft_resnet-101/model_idx_0003/checkpoints/checkpoint-2997",
  "epoch": 9.0,
  "eval_steps": 500,
  "global_step": 2997,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 154.8031005859375,
      "learning_rate": 4.4444444444444447e-05,
      "loss": 3.6286,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5850666666666666,
      "eval_loss": 3.0259337425231934,
      "eval_runtime": 7.7096,
      "eval_samples_per_second": 486.408,
      "eval_steps_per_second": 7.653,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 53.869205474853516,
      "learning_rate": 3.888888888888889e-05,
      "loss": 1.7873,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7698666666666667,
      "eval_loss": 0.9587300419807434,
      "eval_runtime": 7.6265,
      "eval_samples_per_second": 491.706,
      "eval_steps_per_second": 7.736,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 162.40127563476562,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.7565,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8293333333333334,
      "eval_loss": 0.6130377650260925,
      "eval_runtime": 7.6437,
      "eval_samples_per_second": 490.598,
      "eval_steps_per_second": 7.719,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 86.94587707519531,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.5381,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8557333333333333,
      "eval_loss": 0.4922569692134857,
      "eval_runtime": 7.6264,
      "eval_samples_per_second": 491.712,
      "eval_steps_per_second": 7.736,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 118.25404357910156,
      "learning_rate": 2.2222222222222223e-05,
      "loss": 0.4407,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8661333333333333,
      "eval_loss": 0.45935097336769104,
      "eval_runtime": 7.6807,
      "eval_samples_per_second": 488.239,
      "eval_steps_per_second": 7.682,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 164.8728485107422,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 0.366,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8757333333333334,
      "eval_loss": 0.43018534779548645,
      "eval_runtime": 7.6456,
      "eval_samples_per_second": 490.477,
      "eval_steps_per_second": 7.717,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 89.3781509399414,
      "learning_rate": 1.1111111111111112e-05,
      "loss": 0.3247,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.88,
      "eval_loss": 0.408071368932724,
      "eval_runtime": 7.6581,
      "eval_samples_per_second": 489.68,
      "eval_steps_per_second": 7.704,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 115.55562591552734,
      "learning_rate": 5.555555555555556e-06,
      "loss": 0.303,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8786666666666667,
      "eval_loss": 0.40728554129600525,
      "eval_runtime": 7.5946,
      "eval_samples_per_second": 493.773,
      "eval_steps_per_second": 7.769,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 353.60870361328125,
      "learning_rate": 0.0,
      "loss": 0.2825,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8906666666666667,
      "eval_loss": 0.39179283380508423,
      "eval_runtime": 7.66,
      "eval_samples_per_second": 489.559,
      "eval_steps_per_second": 7.702,
      "step": 2997
    }
  ],
  "logging_steps": 500,
  "max_steps": 2997,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 7.3587863155968e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}