trainer_state.json

{
  "best_metric": 0.8845333333333333,
  "best_model_checkpoint": "./resnet_finetuned_models_dataset/CIFAR100/50_from_100/microsoft_resnet-101/model_idx_0304/checkpoints/checkpoint-2331",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 2331,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 71.2537841796875,
      "learning_rate": 4.849231551964771e-05,
      "loss": 3.5617,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6096,
      "eval_loss": 2.648347854614258,
      "eval_runtime": 8.9587,
      "eval_samples_per_second": 418.587,
      "eval_steps_per_second": 6.586,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 272.6517333984375,
      "learning_rate": 4.415111107797445e-05,
      "loss": 1.388,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8029333333333334,
      "eval_loss": 0.7555336952209473,
      "eval_runtime": 8.6533,
      "eval_samples_per_second": 433.363,
      "eval_steps_per_second": 6.818,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 188.6832733154297,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.6569,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8474666666666667,
      "eval_loss": 0.522769570350647,
      "eval_runtime": 8.8783,
      "eval_samples_per_second": 422.376,
      "eval_steps_per_second": 6.645,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 79.6083984375,
      "learning_rate": 2.9341204441673266e-05,
      "loss": 0.4775,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8669333333333333,
      "eval_loss": 0.4387284517288208,
      "eval_runtime": 8.9759,
      "eval_samples_per_second": 417.784,
      "eval_steps_per_second": 6.573,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 149.0272979736328,
      "learning_rate": 2.0658795558326743e-05,
      "loss": 0.3789,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8789333333333333,
      "eval_loss": 0.40560516715049744,
      "eval_runtime": 8.0156,
      "eval_samples_per_second": 467.84,
      "eval_steps_per_second": 7.361,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 194.63626098632812,
      "learning_rate": 1.2500000000000006e-05,
      "loss": 0.3369,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.872,
      "eval_loss": 0.418069988489151,
      "eval_runtime": 8.545,
      "eval_samples_per_second": 438.851,
      "eval_steps_per_second": 6.905,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 73.81818389892578,
      "learning_rate": 5.848888922025553e-06,
      "loss": 0.3056,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8845333333333333,
      "eval_loss": 0.382536381483078,
      "eval_runtime": 8.3098,
      "eval_samples_per_second": 451.273,
      "eval_steps_per_second": 7.1,
      "step": 2331
    }
  ],
  "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": false
      },
      "attributes": {}
    }
  },
  "total_flos": 5.7235004676864e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}