trainer_state.json

{
  "best_global_step": 1467,
  "best_metric": 0.4698332323967337,
  "best_model_checkpoint": "./results/vit5_sft_lora/checkpoint-1467",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 1467,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_loss": 1.3446210622787476,
      "eval_rouge1": 0.731312627714185,
      "eval_rouge2": 0.41556078328954726,
      "eval_rougeL": 0.447114807653285,
      "eval_runtime": 548.2296,
      "eval_samples_per_second": 0.912,
      "eval_steps_per_second": 0.038,
      "step": 489
    },
    {
      "epoch": 1.0224948875255624,
      "grad_norm": 0.6218709349632263,
      "learning_rate": 0.00036666666666666667,
      "loss": 1.7366,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_loss": 1.2575373649597168,
      "eval_rouge1": 0.7501981967742918,
      "eval_rouge2": 0.4449335670061387,
      "eval_rougeL": 0.46616363509873965,
      "eval_runtime": 480.4151,
      "eval_samples_per_second": 1.041,
      "eval_steps_per_second": 0.044,
      "step": 978
    },
    {
      "epoch": 2.044989775051125,
      "grad_norm": 0.6248633861541748,
      "learning_rate": 0.00017727272727272728,
      "loss": 1.4243,
      "step": 1000
    },
    {
      "epoch": 3.0,
      "eval_loss": 1.2363722324371338,
      "eval_rouge1": 0.7520578020362192,
      "eval_rouge2": 0.4468916451304603,
      "eval_rougeL": 0.4698332323967337,
      "eval_runtime": 492.243,
      "eval_samples_per_second": 1.016,
      "eval_steps_per_second": 0.043,
      "step": 1467
    }
  ],
  "logging_steps": 500,
  "max_steps": 1467,
  "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": 5.851496061075456e+16,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}