run-b613dq96/checkpoint-150/trainer_state.json

{
  "best_metric": 0.8192771084337349,
  "best_model_checkpoint": "distilbert-base-multilingual-cased-hyper-matt/run-b613dq96/checkpoint-150",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 150,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2,
      "grad_norm": 1.155951738357544,
      "learning_rate": 4.210286215541667e-05,
      "loss": 0.5868,
      "step": 10
    },
    {
      "epoch": 0.4,
      "grad_norm": 1.9371240139007568,
      "learning_rate": 3.90955148586012e-05,
      "loss": 0.3773,
      "step": 20
    },
    {
      "epoch": 0.6,
      "grad_norm": 1.3020845651626587,
      "learning_rate": 3.608816756178572e-05,
      "loss": 0.393,
      "step": 30
    },
    {
      "epoch": 0.8,
      "grad_norm": 4.894543170928955,
      "learning_rate": 3.308082026497024e-05,
      "loss": 0.4164,
      "step": 40
    },
    {
      "epoch": 1.0,
      "grad_norm": 5.350757598876953,
      "learning_rate": 3.0073472968154765e-05,
      "loss": 0.3851,
      "step": 50
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8825,
      "eval_f1": 0.7892376681614349,
      "eval_loss": 0.36936432123184204,
      "eval_precision": 0.88,
      "eval_recall": 0.7154471544715447,
      "eval_runtime": 1.5306,
      "eval_samples_per_second": 261.342,
      "eval_steps_per_second": 16.334,
      "step": 50
    },
    {
      "epoch": 1.2,
      "grad_norm": 3.0585944652557373,
      "learning_rate": 2.706612567133929e-05,
      "loss": 0.2862,
      "step": 60
    },
    {
      "epoch": 1.4,
      "grad_norm": 5.365485191345215,
      "learning_rate": 2.4058778374523814e-05,
      "loss": 0.2794,
      "step": 70
    },
    {
      "epoch": 1.6,
      "grad_norm": 2.454425096511841,
      "learning_rate": 2.1051431077708335e-05,
      "loss": 0.298,
      "step": 80
    },
    {
      "epoch": 1.8,
      "grad_norm": 3.8099751472473145,
      "learning_rate": 1.804408378089286e-05,
      "loss": 0.2244,
      "step": 90
    },
    {
      "epoch": 2.0,
      "grad_norm": 3.113168478012085,
      "learning_rate": 1.5036736484077383e-05,
      "loss": 0.2519,
      "step": 100
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.89,
      "eval_f1": 0.8151260504201681,
      "eval_loss": 0.28030890226364136,
      "eval_precision": 0.8434782608695652,
      "eval_recall": 0.7886178861788617,
      "eval_runtime": 1.5211,
      "eval_samples_per_second": 262.968,
      "eval_steps_per_second": 16.435,
      "step": 100
    },
    {
      "epoch": 2.2,
      "grad_norm": 2.4218947887420654,
      "learning_rate": 1.2029389187261907e-05,
      "loss": 0.2,
      "step": 110
    },
    {
      "epoch": 2.4,
      "grad_norm": 4.014515399932861,
      "learning_rate": 9.02204189044643e-06,
      "loss": 0.2247,
      "step": 120
    },
    {
      "epoch": 2.6,
      "grad_norm": 1.9129769802093506,
      "learning_rate": 6.0146945936309535e-06,
      "loss": 0.178,
      "step": 130
    },
    {
      "epoch": 2.8,
      "grad_norm": 4.968961715698242,
      "learning_rate": 3.0073472968154768e-06,
      "loss": 0.2099,
      "step": 140
    },
    {
      "epoch": 3.0,
      "grad_norm": 5.08315896987915,
      "learning_rate": 0.0,
      "loss": 0.1404,
      "step": 150
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8875,
      "eval_f1": 0.8192771084337349,
      "eval_loss": 0.28243759274482727,
      "eval_precision": 0.8095238095238095,
      "eval_recall": 0.8292682926829268,
      "eval_runtime": 1.5783,
      "eval_samples_per_second": 253.434,
      "eval_steps_per_second": 15.84,
      "step": 150
    }
  ],
  "logging_steps": 10,
  "max_steps": 150,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "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": 635446111352832.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "_wandb": {},
    "assignments": {},
    "learning_rate": 4.511020945223215e-05,
    "metric": "eval/loss",
    "num_train_epochs": 3,
    "per_device_train_batch_size": 32,
    "seed": 24
  }
}