checkpoint-3711/trainer_state.json

{
  "best_global_step": 3711,
  "best_metric": 0.9990800023078918,
  "best_model_checkpoint": "/my_model_output/rumodernbert_solyanka/checkpoint-3711",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 3711,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.15979520344920506,
      "grad_norm": 3.437915086746216,
      "learning_rate": 8.404742656965777e-07,
      "loss": 0.4866,
      "step": 593
    },
    {
      "epoch": 0.3195904068984101,
      "grad_norm": 2.164642095565796,
      "learning_rate": 6.806790622473725e-07,
      "loss": 0.0792,
      "step": 1186
    },
    {
      "epoch": 0.4793856103476152,
      "grad_norm": 2.1435794830322266,
      "learning_rate": 5.208838587981676e-07,
      "loss": 0.0436,
      "step": 1779
    },
    {
      "epoch": 0.6391808137968202,
      "grad_norm": 1.1813297271728516,
      "learning_rate": 3.6108865534896254e-07,
      "loss": 0.0355,
      "step": 2372
    },
    {
      "epoch": 0.7989760172460253,
      "grad_norm": 1.226506233215332,
      "learning_rate": 2.0129345189975746e-07,
      "loss": 0.0324,
      "step": 2965
    },
    {
      "epoch": 0.9587712206952304,
      "grad_norm": 2.737544059753418,
      "learning_rate": 4.149824845055241e-08,
      "loss": 0.0304,
      "step": 3558
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.030859705060720444,
      "eval_runtime": 1596.6388,
      "eval_samples_per_second": 62.632,
      "eval_steps_per_second": 0.245,
      "eval_triplet_eval_cosine_accuracy": 0.9990800023078918,
      "step": 3711
    }
  ],
  "logging_steps": 593,
  "max_steps": 3711,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "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": 0.0,
  "train_batch_size": 512,
  "trial_name": null,
  "trial_params": null
}