trainer_state.json

{
  "best_metric": 0.8594666666666667,
  "best_model_checkpoint": "./resnet_finetuned_models_dataset/CIFAR100/50_from_100/microsoft_resnet-101/model_idx_0240/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": 142.3601837158203,
      "learning_rate": 2.9095389311788626e-05,
      "loss": 3.8153,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.47173333333333334,
      "eval_loss": 3.5911364555358887,
      "eval_runtime": 8.2334,
      "eval_samples_per_second": 455.462,
      "eval_steps_per_second": 7.166,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 354.9921569824219,
      "learning_rate": 2.649066664678467e-05,
      "loss": 3.0008,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6488,
      "eval_loss": 2.329737424850464,
      "eval_runtime": 7.7211,
      "eval_samples_per_second": 485.684,
      "eval_steps_per_second": 7.641,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 401.3633117675781,
      "learning_rate": 2.25e-05,
      "loss": 1.5389,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7725333333333333,
      "eval_loss": 1.0480810403823853,
      "eval_runtime": 8.1479,
      "eval_samples_per_second": 460.239,
      "eval_steps_per_second": 7.241,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 147.02285766601562,
      "learning_rate": 1.760472266500396e-05,
      "loss": 0.8656,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8248,
      "eval_loss": 0.7041622400283813,
      "eval_runtime": 9.1863,
      "eval_samples_per_second": 408.215,
      "eval_steps_per_second": 6.423,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 151.62693786621094,
      "learning_rate": 1.2395277334996045e-05,
      "loss": 0.6633,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.84,
      "eval_loss": 0.5961145758628845,
      "eval_runtime": 8.6558,
      "eval_samples_per_second": 433.234,
      "eval_steps_per_second": 6.816,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 199.26771545410156,
      "learning_rate": 7.500000000000004e-06,
      "loss": 0.5675,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8469333333333333,
      "eval_loss": 0.5333606004714966,
      "eval_runtime": 8.1649,
      "eval_samples_per_second": 459.283,
      "eval_steps_per_second": 7.226,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 216.65966796875,
      "learning_rate": 3.5093333532153316e-06,
      "loss": 0.5202,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8570666666666666,
      "eval_loss": 0.5124450325965881,
      "eval_runtime": 8.0789,
      "eval_samples_per_second": 464.17,
      "eval_steps_per_second": 7.303,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 255.72842407226562,
      "learning_rate": 9.046106882113753e-07,
      "loss": 0.4881,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8570666666666666,
      "eval_loss": 0.5093125104904175,
      "eval_runtime": 8.077,
      "eval_samples_per_second": 464.282,
      "eval_steps_per_second": 7.305,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 214.293701171875,
      "learning_rate": 0.0,
      "loss": 0.498,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8594666666666667,
      "eval_loss": 0.4966716468334198,
      "eval_runtime": 7.9789,
      "eval_samples_per_second": 469.992,
      "eval_steps_per_second": 7.395,
      "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
}