checkpoint-3752/trainer_state.json

{
  "best_global_step": 3752,
  "best_metric": 0.10899118334054947,
  "best_model_checkpoint": "metaclip-2-image-classification/checkpoint-3752",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 3752,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5330490405117271,
      "grad_norm": 104.10272979736328,
      "learning_rate": 1.7574284170718533e-05,
      "loss": 0.8655,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.92635,
      "eval_loss": 0.2205629199743271,
      "eval_model_preparation_time": 0.0039,
      "eval_runtime": 90.292,
      "eval_samples_per_second": 221.503,
      "eval_steps_per_second": 27.688,
      "step": 938
    },
    {
      "epoch": 1.0660980810234542,
      "grad_norm": 53.17558670043945,
      "learning_rate": 1.4873041599135603e-05,
      "loss": 0.4234,
      "step": 1000
    },
    {
      "epoch": 1.5991471215351813,
      "grad_norm": 17.705827713012695,
      "learning_rate": 1.2171799027552675e-05,
      "loss": 0.3154,
      "step": 1500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9459,
      "eval_loss": 0.15963459014892578,
      "eval_model_preparation_time": 0.0039,
      "eval_runtime": 87.0564,
      "eval_samples_per_second": 229.736,
      "eval_steps_per_second": 28.717,
      "step": 1876
    },
    {
      "epoch": 2.1321961620469083,
      "grad_norm": 21.37250518798828,
      "learning_rate": 9.470556455969747e-06,
      "loss": 0.2481,
      "step": 2000
    },
    {
      "epoch": 2.6652452025586353,
      "grad_norm": 32.63026809692383,
      "learning_rate": 6.769313884386819e-06,
      "loss": 0.2002,
      "step": 2500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9582,
      "eval_loss": 0.1238662451505661,
      "eval_model_preparation_time": 0.0039,
      "eval_runtime": 92.3095,
      "eval_samples_per_second": 216.662,
      "eval_steps_per_second": 27.083,
      "step": 2814
    },
    {
      "epoch": 3.1982942430703627,
      "grad_norm": 37.74256896972656,
      "learning_rate": 4.06807131280389e-06,
      "loss": 0.1637,
      "step": 3000
    },
    {
      "epoch": 3.7313432835820897,
      "grad_norm": 6.652950763702393,
      "learning_rate": 1.3668287412209618e-06,
      "loss": 0.1253,
      "step": 3500
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.96315,
      "eval_loss": 0.10899118334054947,
      "eval_model_preparation_time": 0.0039,
      "eval_runtime": 87.3412,
      "eval_samples_per_second": 228.987,
      "eval_steps_per_second": 28.623,
      "step": 3752
    }
  ],
  "logging_steps": 500,
  "max_steps": 3752,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "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": 2.34038827008e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}