checkpoint-7089/trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.9983086680761097,
  "eval_steps": 1000,
  "global_step": 7089,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00042319085907744394,
      "grad_norm": 8.33436107635498,
      "learning_rate": 0.0,
      "loss": 4.1437,
      "step": 1
    },
    {
      "epoch": 0.4231908590774439,
      "grad_norm": 4.534207820892334,
      "learning_rate": 2.8641065830721004e-05,
      "loss": 3.6457,
      "step": 1000
    },
    {
      "epoch": 0.4231908590774439,
      "eval_cosine_accuracy": 0.9362708926200867,
      "eval_loss": 0.9829329252243042,
      "eval_runtime": 62.6077,
      "eval_samples_per_second": 151.882,
      "eval_steps_per_second": 0.607,
      "step": 1000
    },
    {
      "epoch": 0.8463817181548878,
      "grad_norm": 8.568726539611816,
      "learning_rate": 2.393887147335423e-05,
      "loss": 2.6898,
      "step": 2000
    },
    {
      "epoch": 0.8463817181548878,
      "eval_cosine_accuracy": 0.9378483295440674,
      "eval_loss": 0.9382773041725159,
      "eval_runtime": 58.9559,
      "eval_samples_per_second": 161.29,
      "eval_steps_per_second": 0.645,
      "step": 2000
    },
    {
      "epoch": 1.2693446088794926,
      "grad_norm": 10.338318824768066,
      "learning_rate": 1.9241379310344826e-05,
      "loss": 2.4448,
      "step": 3000
    },
    {
      "epoch": 1.2693446088794926,
      "eval_cosine_accuracy": 0.9374276995658875,
      "eval_loss": 0.980748176574707,
      "eval_runtime": 61.8716,
      "eval_samples_per_second": 153.689,
      "eval_steps_per_second": 0.614,
      "step": 3000
    },
    {
      "epoch": 1.6921775898520086,
      "grad_norm": 4.84547233581543,
      "learning_rate": 1.4539184952978057e-05,
      "loss": 2.6158,
      "step": 4000
    },
    {
      "epoch": 1.6921775898520086,
      "eval_cosine_accuracy": 0.9422652125358582,
      "eval_loss": 0.9552797675132751,
      "eval_runtime": 60.3552,
      "eval_samples_per_second": 157.551,
      "eval_steps_per_second": 0.63,
      "step": 4000
    },
    {
      "epoch": 2.1150105708245244,
      "grad_norm": 8.14175033569336,
      "learning_rate": 9.836990595611285e-06,
      "loss": 2.5157,
      "step": 5000
    },
    {
      "epoch": 2.1150105708245244,
      "eval_cosine_accuracy": 0.9416342377662659,
      "eval_loss": 0.9649843573570251,
      "eval_runtime": 62.8479,
      "eval_samples_per_second": 151.302,
      "eval_steps_per_second": 0.605,
      "step": 5000
    },
    {
      "epoch": 2.53784355179704,
      "grad_norm": 6.472080230712891,
      "learning_rate": 5.13949843260188e-06,
      "loss": 2.3962,
      "step": 6000
    },
    {
      "epoch": 2.53784355179704,
      "eval_cosine_accuracy": 0.9438426494598389,
      "eval_loss": 0.9474993944168091,
      "eval_runtime": 60.8199,
      "eval_samples_per_second": 156.347,
      "eval_steps_per_second": 0.625,
      "step": 6000
    },
    {
      "epoch": 2.960676532769556,
      "grad_norm": 8.412610054016113,
      "learning_rate": 4.3730407523510976e-07,
      "loss": 2.4768,
      "step": 7000
    },
    {
      "epoch": 2.960676532769556,
      "eval_cosine_accuracy": 0.9443684816360474,
      "eval_loss": 0.9480987191200256,
      "eval_runtime": 61.6608,
      "eval_samples_per_second": 154.215,
      "eval_steps_per_second": 0.616,
      "step": 7000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 7089,
  "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": 0.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}