trainer_state.json

{
  "best_metric": 0.9186666666666666,
  "best_model_checkpoint": "./resnet_finetuned_models_dataset/CIFAR100/50_from_100/microsoft_resnet-101/model_idx_0059/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": 29.152673721313477,
      "learning_rate": 0.0004849231551964771,
      "loss": 1.324,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8365333333333334,
      "eval_loss": 0.568449079990387,
      "eval_runtime": 12.6541,
      "eval_samples_per_second": 296.346,
      "eval_steps_per_second": 4.663,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 17.944917678833008,
      "learning_rate": 0.0004415111107797445,
      "loss": 0.4474,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8637333333333334,
      "eval_loss": 0.4810151755809784,
      "eval_runtime": 12.6159,
      "eval_samples_per_second": 297.244,
      "eval_steps_per_second": 4.677,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 26.663326263427734,
      "learning_rate": 0.000375,
      "loss": 0.2584,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8717333333333334,
      "eval_loss": 0.47679102420806885,
      "eval_runtime": 12.5682,
      "eval_samples_per_second": 298.371,
      "eval_steps_per_second": 4.694,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 24.869760513305664,
      "learning_rate": 0.00029341204441673266,
      "loss": 0.1499,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8845333333333333,
      "eval_loss": 0.431570827960968,
      "eval_runtime": 12.6644,
      "eval_samples_per_second": 296.105,
      "eval_steps_per_second": 4.659,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 15.36279582977295,
      "learning_rate": 0.00020658795558326743,
      "loss": 0.0794,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9010666666666667,
      "eval_loss": 0.39594244956970215,
      "eval_runtime": 12.7746,
      "eval_samples_per_second": 293.552,
      "eval_steps_per_second": 4.619,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 27.67572021484375,
      "learning_rate": 0.00012500000000000006,
      "loss": 0.0519,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9045333333333333,
      "eval_loss": 0.3909560739994049,
      "eval_runtime": 12.7545,
      "eval_samples_per_second": 294.014,
      "eval_steps_per_second": 4.626,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 23.077165603637695,
      "learning_rate": 5.848888922025553e-05,
      "loss": 0.0251,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9122666666666667,
      "eval_loss": 0.3800540268421173,
      "eval_runtime": 12.642,
      "eval_samples_per_second": 296.63,
      "eval_steps_per_second": 4.667,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 25.55421257019043,
      "learning_rate": 1.5076844803522921e-05,
      "loss": 0.0198,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9122666666666667,
      "eval_loss": 0.36937227845191956,
      "eval_runtime": 13.0798,
      "eval_samples_per_second": 286.702,
      "eval_steps_per_second": 4.511,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 18.439367294311523,
      "learning_rate": 0.0,
      "loss": 0.0129,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9186666666666666,
      "eval_loss": 0.3570243716239929,
      "eval_runtime": 12.5426,
      "eval_samples_per_second": 298.98,
      "eval_steps_per_second": 4.704,
      "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
}