run-0/checkpoint-50514/trainer_state.json

{
  "best_metric": 0.8841743119266054,
  "best_model_checkpoint": "Tech-oriented/best_model_bert_uncasedbert-base-uncased-finetuned-sst2/run-0/checkpoint-50514",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 50514,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.25,
      "grad_norm": 19.703554153442383,
      "learning_rate": 6.973163454116781e-07,
      "loss": 0.7045,
      "step": 4209
    },
    {
      "epoch": 0.5,
      "grad_norm": 33.55769729614258,
      "learning_rate": 6.508344788890587e-07,
      "loss": 0.6701,
      "step": 8418
    },
    {
      "epoch": 0.75,
      "grad_norm": 121.4453353881836,
      "learning_rate": 6.043526123664394e-07,
      "loss": 0.566,
      "step": 12627
    },
    {
      "epoch": 1.0,
      "grad_norm": 72.47432708740234,
      "learning_rate": 5.578707458438201e-07,
      "loss": 0.5309,
      "step": 16836
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8715596330275229,
      "eval_loss": 0.3419412672519684,
      "eval_runtime": 2.4269,
      "eval_samples_per_second": 359.309,
      "eval_steps_per_second": 22.663,
      "step": 16838
    },
    {
      "epoch": 1.25,
      "grad_norm": 169.0806884765625,
      "learning_rate": 5.113888793212007e-07,
      "loss": 0.5455,
      "step": 21045
    },
    {
      "epoch": 1.5,
      "grad_norm": 26.04638671875,
      "learning_rate": 4.649070127985814e-07,
      "loss": 0.5548,
      "step": 25254
    },
    {
      "epoch": 1.75,
      "grad_norm": 17.553491592407227,
      "learning_rate": 4.18425146275962e-07,
      "loss": 0.5681,
      "step": 29463
    },
    {
      "epoch": 2.0,
      "grad_norm": 3.3310186862945557,
      "learning_rate": 3.7194327975334265e-07,
      "loss": 0.6066,
      "step": 33672
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8784403669724771,
      "eval_loss": 0.3679516315460205,
      "eval_runtime": 2.4422,
      "eval_samples_per_second": 357.051,
      "eval_steps_per_second": 22.52,
      "step": 33676
    },
    {
      "epoch": 2.25,
      "grad_norm": 60.65003967285156,
      "learning_rate": 3.2546141323072335e-07,
      "loss": 0.5873,
      "step": 37881
    },
    {
      "epoch": 2.5,
      "grad_norm": 179.75804138183594,
      "learning_rate": 2.78979546708104e-07,
      "loss": 0.6103,
      "step": 42090
    },
    {
      "epoch": 2.75,
      "grad_norm": 0.3105064332485199,
      "learning_rate": 2.3249768018548465e-07,
      "loss": 0.6097,
      "step": 46299
    },
    {
      "epoch": 3.0,
      "grad_norm": 22.6956844329834,
      "learning_rate": 1.8601581366286533e-07,
      "loss": 0.5877,
      "step": 50508
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8841743119266054,
      "eval_loss": 0.40006375312805176,
      "eval_runtime": 2.4426,
      "eval_samples_per_second": 357.004,
      "eval_steps_per_second": 22.517,
      "step": 50514
    }
  ],
  "logging_steps": 4209,
  "max_steps": 67352,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "total_flos": 2484012973657680.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 7.437982119342974e-07,
    "num_train_epochs": 4,
    "per_device_train_batch_size": 4,
    "seed": 7,
    "weight_decay": 0.46186751115854463
  }
}