trainer_state.json

{
  "best_metric": 0.7685333333333333,
  "best_model_checkpoint": "./resnet_finetuned_models_dataset/CIFAR100/50_from_100/microsoft_resnet-101/model_idx_0685/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": 110.63288879394531,
      "learning_rate": 2.8514533018536286e-05,
      "loss": 3.8589,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.3181333333333333,
      "eval_loss": 3.759841203689575,
      "eval_runtime": 8.9243,
      "eval_samples_per_second": 420.201,
      "eval_steps_per_second": 6.611,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 109.01708984375,
      "learning_rate": 2.4352347027881003e-05,
      "loss": 3.3089,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5053333333333333,
      "eval_loss": 2.7527120113372803,
      "eval_runtime": 9.0532,
      "eval_samples_per_second": 414.217,
      "eval_steps_per_second": 6.517,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 263.4269104003906,
      "learning_rate": 1.8337814009344716e-05,
      "loss": 2.0056,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6594666666666666,
      "eval_loss": 1.5119948387145996,
      "eval_runtime": 8.7026,
      "eval_samples_per_second": 430.903,
      "eval_steps_per_second": 6.78,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 126.03755950927734,
      "learning_rate": 1.1662185990655285e-05,
      "loss": 1.3266,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7290666666666666,
      "eval_loss": 1.1502372026443481,
      "eval_runtime": 8.5501,
      "eval_samples_per_second": 438.589,
      "eval_steps_per_second": 6.9,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 179.62347412109375,
      "learning_rate": 5.647652972118998e-06,
      "loss": 1.0831,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.752,
      "eval_loss": 1.0169949531555176,
      "eval_runtime": 8.5192,
      "eval_samples_per_second": 440.184,
      "eval_steps_per_second": 6.926,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 194.9832000732422,
      "learning_rate": 1.4854669814637145e-06,
      "loss": 0.9923,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7578666666666667,
      "eval_loss": 0.9954235553741455,
      "eval_runtime": 8.5764,
      "eval_samples_per_second": 437.246,
      "eval_steps_per_second": 6.879,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 349.189453125,
      "learning_rate": 0.0,
      "loss": 0.9666,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7685333333333333,
      "eval_loss": 0.9413397312164307,
      "eval_runtime": 8.591,
      "eval_samples_per_second": 436.503,
      "eval_steps_per_second": 6.868,
      "step": 2331
    }
  ],
  "logging_steps": 500,
  "max_steps": 2331,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "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": 5.7235004676864e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}