trainer_state.json

{
  "best_metric": 0.8333333333333334,
  "best_model_checkpoint": "./models/roberta/checkpoint-56",
  "epoch": 8.0,
  "eval_steps": 500,
  "global_step": 56,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.6466717720031738,
      "learning_rate": 9.666666666666667e-05,
      "loss": 0.1697,
      "step": 7
    },
    {
      "epoch": 1.0,
      "eval_f1_macro": 0.35912698412698413,
      "eval_f1_micro": 0.45,
      "eval_loss": 0.1369188129901886,
      "eval_runtime": 7.5543,
      "eval_samples_per_second": 1.589,
      "eval_steps_per_second": 0.132,
      "step": 7
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.6866563558578491,
      "learning_rate": 9.333333333333334e-05,
      "loss": 0.1247,
      "step": 14
    },
    {
      "epoch": 2.0,
      "eval_f1_macro": 0.49907407407407406,
      "eval_f1_micro": 0.5641025641025641,
      "eval_loss": 0.09602320194244385,
      "eval_runtime": 6.4291,
      "eval_samples_per_second": 1.867,
      "eval_steps_per_second": 0.156,
      "step": 14
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.0547248125076294,
      "learning_rate": 9e-05,
      "loss": 0.0901,
      "step": 21
    },
    {
      "epoch": 3.0,
      "eval_f1_macro": 0.7619047619047619,
      "eval_f1_micro": 0.75,
      "eval_loss": 0.07866408675909042,
      "eval_runtime": 5.9612,
      "eval_samples_per_second": 2.013,
      "eval_steps_per_second": 0.168,
      "step": 21
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.8302876949310303,
      "learning_rate": 8.666666666666667e-05,
      "loss": 0.0673,
      "step": 28
    },
    {
      "epoch": 4.0,
      "eval_f1_macro": 0.7431818181818182,
      "eval_f1_micro": 0.8333333333333334,
      "eval_loss": 0.0431751012802124,
      "eval_runtime": 6.5441,
      "eval_samples_per_second": 1.834,
      "eval_steps_per_second": 0.153,
      "step": 28
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.7726091146469116,
      "learning_rate": 8.333333333333334e-05,
      "loss": 0.0564,
      "step": 35
    },
    {
      "epoch": 5.0,
      "eval_f1_macro": 0.8095238095238096,
      "eval_f1_micro": 0.8333333333333334,
      "eval_loss": 0.04845364764332771,
      "eval_runtime": 6.4361,
      "eval_samples_per_second": 1.864,
      "eval_steps_per_second": 0.155,
      "step": 35
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.5374885201454163,
      "learning_rate": 8e-05,
      "loss": 0.0326,
      "step": 42
    },
    {
      "epoch": 6.0,
      "eval_f1_macro": 0.8095238095238096,
      "eval_f1_micro": 0.967741935483871,
      "eval_loss": 0.024476679041981697,
      "eval_runtime": 5.7698,
      "eval_samples_per_second": 2.08,
      "eval_steps_per_second": 0.173,
      "step": 42
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.8273659348487854,
      "learning_rate": 7.666666666666667e-05,
      "loss": 0.0282,
      "step": 49
    },
    {
      "epoch": 7.0,
      "eval_f1_macro": 0.8148148148148149,
      "eval_f1_micro": 0.9333333333333333,
      "eval_loss": 0.027787813916802406,
      "eval_runtime": 5.897,
      "eval_samples_per_second": 2.035,
      "eval_steps_per_second": 0.17,
      "step": 49
    },
    {
      "epoch": 8.0,
      "grad_norm": 1.7157799005508423,
      "learning_rate": 7.333333333333333e-05,
      "loss": 0.0243,
      "step": 56
    },
    {
      "epoch": 8.0,
      "eval_f1_macro": 0.8333333333333334,
      "eval_f1_micro": 1.0,
      "eval_loss": 0.010664558969438076,
      "eval_runtime": 5.9437,
      "eval_samples_per_second": 2.019,
      "eval_steps_per_second": 0.168,
      "step": 56
    }
  ],
  "logging_steps": 500,
  "max_steps": 210,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 30,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 225231152136192.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}