checkpoint-9830/trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.997966446365023,
  "eval_steps": 1000,
  "global_step": 9830,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.000508646998982706,
      "grad_norm": 4.802672863006592,
      "learning_rate": 0.0,
      "loss": 3.5148,
      "step": 1
    },
    {
      "epoch": 0.508646998982706,
      "grad_norm": 4.325772285461426,
      "learning_rate": 2.994574432010851e-05,
      "loss": 2.3332,
      "step": 1000
    },
    {
      "epoch": 0.508646998982706,
      "eval_cosine_accuracy": 0.9514144659042358,
      "eval_loss": 0.4182649254798889,
      "eval_runtime": 40.3586,
      "eval_samples_per_second": 235.613,
      "eval_steps_per_second": 0.942,
      "step": 1000
    },
    {
      "epoch": 1.0172852058973056,
      "grad_norm": 4.248680114746094,
      "learning_rate": 2.6554764326890472e-05,
      "loss": 1.1848,
      "step": 2000
    },
    {
      "epoch": 1.0172852058973056,
      "eval_cosine_accuracy": 0.9598275423049927,
      "eval_loss": 0.36312881112098694,
      "eval_runtime": 39.402,
      "eval_samples_per_second": 241.333,
      "eval_steps_per_second": 0.964,
      "step": 2000
    },
    {
      "epoch": 1.5256736146415861,
      "grad_norm": 3.5537898540496826,
      "learning_rate": 2.316378433367243e-05,
      "loss": 1.2233,
      "step": 3000
    },
    {
      "epoch": 1.5256736146415861,
      "eval_cosine_accuracy": 0.964454710483551,
      "eval_loss": 0.35020381212234497,
      "eval_runtime": 39.2709,
      "eval_samples_per_second": 242.139,
      "eval_steps_per_second": 0.968,
      "step": 3000
    },
    {
      "epoch": 2.0340620233858666,
      "grad_norm": 3.946258068084717,
      "learning_rate": 1.9772804340454392e-05,
      "loss": 1.0983,
      "step": 4000
    },
    {
      "epoch": 2.0340620233858666,
      "eval_cosine_accuracy": 0.9663476943969727,
      "eval_loss": 0.34889712929725647,
      "eval_runtime": 39.8981,
      "eval_samples_per_second": 238.332,
      "eval_steps_per_second": 0.952,
      "step": 4000
    },
    {
      "epoch": 2.5424504321301473,
      "grad_norm": 4.344015598297119,
      "learning_rate": 1.638182434723635e-05,
      "loss": 0.9919,
      "step": 5000
    },
    {
      "epoch": 2.5424504321301473,
      "eval_cosine_accuracy": 0.9686612486839294,
      "eval_loss": 0.33659711480140686,
      "eval_runtime": 40.7722,
      "eval_samples_per_second": 233.222,
      "eval_steps_per_second": 0.932,
      "step": 5000
    },
    {
      "epoch": 3.050838840874428,
      "grad_norm": 3.2069296836853027,
      "learning_rate": 1.2990844354018311e-05,
      "loss": 0.9487,
      "step": 6000
    },
    {
      "epoch": 3.050838840874428,
      "eval_cosine_accuracy": 0.9696077108383179,
      "eval_loss": 0.3368596136569977,
      "eval_runtime": 40.3843,
      "eval_samples_per_second": 235.463,
      "eval_steps_per_second": 0.941,
      "step": 6000
    },
    {
      "epoch": 3.5592272496187087,
      "grad_norm": 3.514526844024658,
      "learning_rate": 9.606646320786708e-06,
      "loss": 0.8887,
      "step": 7000
    },
    {
      "epoch": 3.5592272496187087,
      "eval_cosine_accuracy": 0.9696077108383179,
      "eval_loss": 0.33110618591308594,
      "eval_runtime": 41.4945,
      "eval_samples_per_second": 229.163,
      "eval_steps_per_second": 0.916,
      "step": 7000
    },
    {
      "epoch": 4.067615658362989,
      "grad_norm": 3.503282070159912,
      "learning_rate": 6.215666327568667e-06,
      "loss": 0.8724,
      "step": 8000
    },
    {
      "epoch": 4.067615658362989,
      "eval_cosine_accuracy": 0.9703438878059387,
      "eval_loss": 0.3360755741596222,
      "eval_runtime": 39.6305,
      "eval_samples_per_second": 239.941,
      "eval_steps_per_second": 0.959,
      "step": 8000
    },
    {
      "epoch": 4.57600406710727,
      "grad_norm": 3.249732732772827,
      "learning_rate": 2.8280773143438456e-06,
      "loss": 0.8366,
      "step": 9000
    },
    {
      "epoch": 4.57600406710727,
      "eval_cosine_accuracy": 0.9704490303993225,
      "eval_loss": 0.33363083004951477,
      "eval_runtime": 40.7633,
      "eval_samples_per_second": 233.273,
      "eval_steps_per_second": 0.932,
      "step": 9000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 9830,
  "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
}