trainer_state.json

{
  "best_metric": 0.9266666666666666,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0904/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": Infinity,
      "learning_rate": 2.9106113962682575e-05,
      "loss": 0.7916,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8773333333333333,
      "eval_loss": 0.38078323006629944,
      "eval_runtime": 6.8238,
      "eval_samples_per_second": 549.546,
      "eval_steps_per_second": 8.646,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 206.15118408203125,
      "learning_rate": 2.6510855361961218e-05,
      "loss": 0.1563,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8986666666666666,
      "eval_loss": 0.3265850841999054,
      "eval_runtime": 6.5249,
      "eval_samples_per_second": 574.718,
      "eval_steps_per_second": 9.042,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.19263404607772827,
      "learning_rate": 2.2527217722455156e-05,
      "loss": 0.0575,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9013333333333333,
      "eval_loss": 0.36817702651023865,
      "eval_runtime": 6.5047,
      "eval_samples_per_second": 576.51,
      "eval_steps_per_second": 9.07,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.00460640387609601,
      "learning_rate": 1.763568653571882e-05,
      "loss": 0.021,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9133333333333333,
      "eval_loss": 0.36653801798820496,
      "eval_runtime": 6.6575,
      "eval_samples_per_second": 563.278,
      "eval_steps_per_second": 8.862,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.0072155240923166275,
      "learning_rate": 1.2426252654184347e-05,
      "loss": 0.0047,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9242666666666667,
      "eval_loss": 0.3330228626728058,
      "eval_runtime": 6.499,
      "eval_samples_per_second": 577.012,
      "eval_steps_per_second": 9.078,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.03789821267127991,
      "learning_rate": 7.5272506870206094e-06,
      "loss": 0.0007,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9237333333333333,
      "eval_loss": 0.34245893359184265,
      "eval_runtime": 6.4761,
      "eval_samples_per_second": 579.052,
      "eval_steps_per_second": 9.11,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.001275436021387577,
      "learning_rate": 3.5295725730362483e-06,
      "loss": 0.0004,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9264,
      "eval_loss": 0.3385518789291382,
      "eval_runtime": 6.4586,
      "eval_samples_per_second": 580.619,
      "eval_steps_per_second": 9.135,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.005607640836387873,
      "learning_rate": 9.153972922231291e-07,
      "loss": 0.0003,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9264,
      "eval_loss": 0.33987343311309814,
      "eval_runtime": 6.4462,
      "eval_samples_per_second": 581.742,
      "eval_steps_per_second": 9.153,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.0010035211453214288,
      "learning_rate": 3.2964565453985627e-11,
      "loss": 0.0001,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9266666666666666,
      "eval_loss": 0.3401177227497101,
      "eval_runtime": 6.7935,
      "eval_samples_per_second": 552.001,
      "eval_steps_per_second": 8.685,
      "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": 1.482671885603328e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}