checkpoint-10825/trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "eval_steps": 1000,
  "global_step": 10825,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00046189376443418013,
      "grad_norm": 7.368561744689941,
      "learning_rate": 0.0,
      "loss": 3.6183,
      "step": 1
    },
    {
      "epoch": 0.4618937644341801,
      "eval_cosine_accuracy": 0.9518148303031921,
      "eval_loss": 1.2476075887680054,
      "eval_runtime": 21.6936,
      "eval_samples_per_second": 438.147,
      "eval_steps_per_second": 1.752,
      "step": 1000
    },
    {
      "epoch": 0.9237875288683602,
      "eval_cosine_accuracy": 0.9532877206802368,
      "eval_loss": 1.1978851556777954,
      "eval_runtime": 21.447,
      "eval_samples_per_second": 443.185,
      "eval_steps_per_second": 1.772,
      "step": 2000
    },
    {
      "epoch": 1.0,
      "grad_norm": 18.16141128540039,
      "learning_rate": 1.7780970953505083e-05,
      "loss": 2.4107,
      "step": 2165
    },
    {
      "epoch": 1.3856812933025404,
      "eval_cosine_accuracy": 0.9571804404258728,
      "eval_loss": 1.1492724418640137,
      "eval_runtime": 21.5248,
      "eval_samples_per_second": 441.583,
      "eval_steps_per_second": 1.765,
      "step": 3000
    },
    {
      "epoch": 1.8475750577367207,
      "eval_cosine_accuracy": 0.9597054123878479,
      "eval_loss": 1.1209757328033447,
      "eval_runtime": 21.5534,
      "eval_samples_per_second": 440.998,
      "eval_steps_per_second": 1.763,
      "step": 4000
    },
    {
      "epoch": 2.0,
      "grad_norm": 17.87831687927246,
      "learning_rate": 1.3338807348865855e-05,
      "loss": 1.8868,
      "step": 4330
    },
    {
      "epoch": 2.3094688221709005,
      "eval_cosine_accuracy": 0.9584429264068604,
      "eval_loss": 1.1083347797393799,
      "eval_runtime": 21.4748,
      "eval_samples_per_second": 442.613,
      "eval_steps_per_second": 1.77,
      "step": 5000
    },
    {
      "epoch": 2.771362586605081,
      "eval_cosine_accuracy": 0.9596002101898193,
      "eval_loss": 1.0890077352523804,
      "eval_runtime": 22.6436,
      "eval_samples_per_second": 419.764,
      "eval_steps_per_second": 1.678,
      "step": 6000
    },
    {
      "epoch": 3.0,
      "grad_norm": 19.0833797454834,
      "learning_rate": 8.898696500051319e-06,
      "loss": 1.7202,
      "step": 6495
    },
    {
      "epoch": 3.233256351039261,
      "eval_cosine_accuracy": 0.9590741991996765,
      "eval_loss": 1.083592414855957,
      "eval_runtime": 21.6954,
      "eval_samples_per_second": 438.112,
      "eval_steps_per_second": 1.752,
      "step": 7000
    },
    {
      "epoch": 3.695150115473441,
      "eval_cosine_accuracy": 0.9600210189819336,
      "eval_loss": 1.0752122402191162,
      "eval_runtime": 22.018,
      "eval_samples_per_second": 431.692,
      "eval_steps_per_second": 1.726,
      "step": 8000
    },
    {
      "epoch": 4.0,
      "grad_norm": 19.486677169799805,
      "learning_rate": 4.4544801395873965e-06,
      "loss": 1.6355,
      "step": 8660
    },
    {
      "epoch": 4.157043879907621,
      "eval_cosine_accuracy": 0.9589689373970032,
      "eval_loss": 1.0678318738937378,
      "eval_runtime": 21.4878,
      "eval_samples_per_second": 442.343,
      "eval_steps_per_second": 1.768,
      "step": 9000
    },
    {
      "epoch": 4.618937644341801,
      "eval_cosine_accuracy": 0.9594950079917908,
      "eval_loss": 1.070329189300537,
      "eval_runtime": 21.5061,
      "eval_samples_per_second": 441.967,
      "eval_steps_per_second": 1.767,
      "step": 10000
    },
    {
      "epoch": 5.0,
      "grad_norm": 14.038162231445312,
      "learning_rate": 1.2316534948167918e-08,
      "loss": 1.5986,
      "step": 10825
    }
  ],
  "logging_steps": 500,
  "max_steps": 10825,
  "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": 0.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}