trainer_state.json

{
  "best_global_step": 540,
  "best_metric": 0.01361137069761753,
  "best_model_checkpoint": "./camembert-secabilite-regressor/checkpoint-540",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 540,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.46296296296296297,
      "grad_norm": 0.0614665187895298,
      "learning_rate": 2.7277777777777777e-05,
      "loss": 0.0155,
      "step": 50
    },
    {
      "epoch": 0.9259259259259259,
      "grad_norm": 0.04598488286137581,
      "learning_rate": 2.45e-05,
      "loss": 0.0144,
      "step": 100
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.014625671319663525,
      "eval_mae": 0.07523933053016663,
      "eval_model_preparation_time": 0.0012,
      "eval_mse": 0.014703812077641487,
      "eval_runtime": 30.6564,
      "eval_samples_per_second": 28.151,
      "eval_steps_per_second": 3.523,
      "step": 108
    },
    {
      "epoch": 1.3888888888888888,
      "grad_norm": 0.057490844279527664,
      "learning_rate": 2.1722222222222225e-05,
      "loss": 0.0133,
      "step": 150
    },
    {
      "epoch": 1.8518518518518519,
      "grad_norm": 0.05683305859565735,
      "learning_rate": 1.8944444444444447e-05,
      "loss": 0.0133,
      "step": 200
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.014103018678724766,
      "eval_mae": 0.07052576541900635,
      "eval_model_preparation_time": 0.0012,
      "eval_mse": 0.014136598445475101,
      "eval_runtime": 31.5256,
      "eval_samples_per_second": 27.375,
      "eval_steps_per_second": 3.426,
      "step": 216
    },
    {
      "epoch": 2.314814814814815,
      "grad_norm": 0.049927715212106705,
      "learning_rate": 1.6166666666666665e-05,
      "loss": 0.0126,
      "step": 250
    },
    {
      "epoch": 2.7777777777777777,
      "grad_norm": 0.05183377489447594,
      "learning_rate": 1.338888888888889e-05,
      "loss": 0.0122,
      "step": 300
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.013754396699368954,
      "eval_mae": 0.06792013347148895,
      "eval_model_preparation_time": 0.0012,
      "eval_mse": 0.013822735287249088,
      "eval_runtime": 30.5607,
      "eval_samples_per_second": 28.239,
      "eval_steps_per_second": 3.534,
      "step": 324
    },
    {
      "epoch": 3.240740740740741,
      "grad_norm": 0.08390484005212784,
      "learning_rate": 1.0611111111111111e-05,
      "loss": 0.0116,
      "step": 350
    },
    {
      "epoch": 3.7037037037037037,
      "grad_norm": 0.05707550421357155,
      "learning_rate": 7.833333333333333e-06,
      "loss": 0.0119,
      "step": 400
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.01372485887259245,
      "eval_mae": 0.06702560186386108,
      "eval_model_preparation_time": 0.0012,
      "eval_mse": 0.013811548240482807,
      "eval_runtime": 30.765,
      "eval_samples_per_second": 28.051,
      "eval_steps_per_second": 3.51,
      "step": 432
    },
    {
      "epoch": 4.166666666666667,
      "grad_norm": 0.07036450505256653,
      "learning_rate": 5.0555555555555555e-06,
      "loss": 0.0115,
      "step": 450
    },
    {
      "epoch": 4.62962962962963,
      "grad_norm": 0.04305871203541756,
      "learning_rate": 2.2777777777777776e-06,
      "loss": 0.0119,
      "step": 500
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.013537541963160038,
      "eval_mae": 0.06590602546930313,
      "eval_model_preparation_time": 0.0012,
      "eval_mse": 0.01361137069761753,
      "eval_runtime": 30.7943,
      "eval_samples_per_second": 28.025,
      "eval_steps_per_second": 3.507,
      "step": 540
    },
    {
      "epoch": 5.0,
      "step": 540,
      "total_flos": 4537309410647040.0,
      "train_loss": 0.012715035952903606,
      "train_runtime": 5221.3545,
      "train_samples_per_second": 3.303,
      "train_steps_per_second": 0.103
    }
  ],
  "logging_steps": 50,
  "max_steps": 540,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "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": 4537309410647040.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}