checkpoint-6495/trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 6495,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00046189376443418013,
      "grad_norm": 12.800962448120117,
      "learning_rate": 0.0,
      "loss": 4.0583,
      "step": 1
    },
    {
      "epoch": 0.23094688221709006,
      "eval_cosine_accuracy": 0.9421356916427612,
      "eval_loss": 1.4610942602157593,
      "eval_runtime": 21.999,
      "eval_samples_per_second": 432.064,
      "eval_steps_per_second": 1.727,
      "step": 500
    },
    {
      "epoch": 0.4618937644341801,
      "eval_cosine_accuracy": 0.9457128047943115,
      "eval_loss": 1.3612146377563477,
      "eval_runtime": 21.8898,
      "eval_samples_per_second": 434.22,
      "eval_steps_per_second": 1.736,
      "step": 1000
    },
    {
      "epoch": 0.6928406466512702,
      "eval_cosine_accuracy": 0.9528669118881226,
      "eval_loss": 1.288307547569275,
      "eval_runtime": 21.8186,
      "eval_samples_per_second": 435.638,
      "eval_steps_per_second": 1.742,
      "step": 1500
    },
    {
      "epoch": 0.9237875288683602,
      "eval_cosine_accuracy": 0.9522356390953064,
      "eval_loss": 1.2684112787246704,
      "eval_runtime": 21.8066,
      "eval_samples_per_second": 435.877,
      "eval_steps_per_second": 1.743,
      "step": 2000
    },
    {
      "epoch": 1.0,
      "grad_norm": 18.849971771240234,
      "learning_rate": 1.6664098613251156e-05,
      "loss": 2.6124,
      "step": 2165
    },
    {
      "epoch": 1.1547344110854503,
      "eval_cosine_accuracy": 0.9541293978691101,
      "eval_loss": 1.2559542655944824,
      "eval_runtime": 21.9828,
      "eval_samples_per_second": 432.383,
      "eval_steps_per_second": 1.729,
      "step": 2500
    },
    {
      "epoch": 1.3856812933025404,
      "eval_cosine_accuracy": 0.956233561038971,
      "eval_loss": 1.1885266304016113,
      "eval_runtime": 23.3704,
      "eval_samples_per_second": 406.71,
      "eval_steps_per_second": 1.626,
      "step": 3000
    },
    {
      "epoch": 1.6166281755196303,
      "eval_cosine_accuracy": 0.9557075500488281,
      "eval_loss": 1.187910556793213,
      "eval_runtime": 21.9291,
      "eval_samples_per_second": 433.442,
      "eval_steps_per_second": 1.733,
      "step": 3500
    },
    {
      "epoch": 1.8475750577367207,
      "eval_cosine_accuracy": 0.9580221176147461,
      "eval_loss": 1.1555284261703491,
      "eval_runtime": 21.9378,
      "eval_samples_per_second": 433.27,
      "eval_steps_per_second": 1.732,
      "step": 4000
    },
    {
      "epoch": 2.0,
      "grad_norm": 17.494279861450195,
      "learning_rate": 1.1120800205180818e-05,
      "loss": 1.986,
      "step": 4330
    },
    {
      "epoch": 2.0785219399538106,
      "eval_cosine_accuracy": 0.9582325220108032,
      "eval_loss": 1.154712438583374,
      "eval_runtime": 22.0846,
      "eval_samples_per_second": 430.391,
      "eval_steps_per_second": 1.721,
      "step": 4500
    },
    {
      "epoch": 2.3094688221709005,
      "eval_cosine_accuracy": 0.9584429264068604,
      "eval_loss": 1.145609974861145,
      "eval_runtime": 22.0404,
      "eval_samples_per_second": 431.254,
      "eval_steps_per_second": 1.724,
      "step": 5000
    },
    {
      "epoch": 2.540415704387991,
      "eval_cosine_accuracy": 0.9585481286048889,
      "eval_loss": 1.135787844657898,
      "eval_runtime": 23.3896,
      "eval_samples_per_second": 406.377,
      "eval_steps_per_second": 1.625,
      "step": 5500
    },
    {
      "epoch": 2.771362586605081,
      "eval_cosine_accuracy": 0.9596002101898193,
      "eval_loss": 1.1278640031814575,
      "eval_runtime": 21.8218,
      "eval_samples_per_second": 435.573,
      "eval_steps_per_second": 1.741,
      "step": 6000
    },
    {
      "epoch": 3.0,
      "grad_norm": 19.504749298095703,
      "learning_rate": 2.0518081559374197e-08,
      "loss": 1.8005,
      "step": 6495
    }
  ],
  "logging_steps": 500,
  "max_steps": 6495,
  "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
}