run-21/checkpoint-960/trainer_state.json

{
  "best_metric": 0.8189823874755382,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-21/checkpoint-576",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 960,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 5.036266803741455,
      "learning_rate": 7.741005485591573e-05,
      "loss": 0.556,
      "step": 192
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.786692759295499,
      "eval_f1": 0.8080985915492958,
      "eval_loss": 0.4501095712184906,
      "eval_mcc": 0.5882099387037968,
      "eval_precision": 0.7344,
      "eval_recall": 0.898238747553816,
      "eval_runtime": 67.0331,
      "eval_samples_per_second": 15.246,
      "eval_steps_per_second": 0.477,
      "step": 192
    },
    {
      "epoch": 2.0,
      "grad_norm": 7.937936305999756,
      "learning_rate": 5.8057541141936786e-05,
      "loss": 0.4586,
      "step": 384
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8082191780821918,
      "eval_f1": 0.8161350844277674,
      "eval_loss": 0.42737704515457153,
      "eval_mcc": 0.6187363370477468,
      "eval_precision": 0.7837837837837838,
      "eval_recall": 0.8512720156555773,
      "eval_runtime": 66.4748,
      "eval_samples_per_second": 15.374,
      "eval_steps_per_second": 0.481,
      "step": 384
    },
    {
      "epoch": 3.0,
      "grad_norm": 4.516488075256348,
      "learning_rate": 3.8705027427957864e-05,
      "loss": 0.4308,
      "step": 576
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8189823874755382,
      "eval_f1": 0.8291782086795938,
      "eval_loss": 0.4247056245803833,
      "eval_mcc": 0.6425594703081364,
      "eval_precision": 0.784965034965035,
      "eval_recall": 0.8786692759295499,
      "eval_runtime": 66.8952,
      "eval_samples_per_second": 15.278,
      "eval_steps_per_second": 0.478,
      "step": 576
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.640265941619873,
      "learning_rate": 1.9352513713978932e-05,
      "loss": 0.4242,
      "step": 768
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8111545988258317,
      "eval_f1": 0.8214616096207216,
      "eval_loss": 0.42209360003471375,
      "eval_mcc": 0.6264991336111037,
      "eval_precision": 0.7789473684210526,
      "eval_recall": 0.8688845401174168,
      "eval_runtime": 67.1893,
      "eval_samples_per_second": 15.211,
      "eval_steps_per_second": 0.476,
      "step": 768
    },
    {
      "epoch": 5.0,
      "grad_norm": 7.009722709655762,
      "learning_rate": 0.0,
      "loss": 0.419,
      "step": 960
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8180039138943248,
      "eval_f1": 0.8342245989304813,
      "eval_loss": 0.42406341433525085,
      "eval_mcc": 0.6485475826584332,
      "eval_precision": 0.7659574468085106,
      "eval_recall": 0.9158512720156555,
      "eval_runtime": 66.6719,
      "eval_samples_per_second": 15.329,
      "eval_steps_per_second": 0.48,
      "step": 960
    }
  ],
  "logging_steps": 500,
  "max_steps": 960,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 1178475986400.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.8653499588119579,
    "learning_rate": 9.676256856989465e-05,
    "num_train_epochs": 5,
    "per_device_train_batch_size": 16,
    "temperature": 45
  }
}