{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.996376811594203,
  "eval_steps": 1000,
  "global_step": 16548,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00036258158085569254,
      "grad_norm": 7.228061676025391,
      "learning_rate": 0.0,
      "loss": 4.1707,
      "step": 1
    },
    {
      "epoch": 0.36258158085569253,
      "grad_norm": 5.754242897033691,
      "learning_rate": 1.5090634441087614e-05,
      "loss": 3.3117,
      "step": 1000
    },
    {
      "epoch": 0.36258158085569253,
      "eval_cosine_accuracy": 0.9497318267822266,
      "eval_loss": 0.5286146402359009,
      "eval_runtime": 34.6121,
      "eval_samples_per_second": 274.73,
      "eval_steps_per_second": 1.098,
      "step": 1000
    },
    {
      "epoch": 0.7251631617113851,
      "grad_norm": 7.097472190856934,
      "learning_rate": 3.01963746223565e-05,
      "loss": 1.9469,
      "step": 2000
    },
    {
      "epoch": 0.7251631617113851,
      "eval_cosine_accuracy": 0.9579346179962158,
      "eval_loss": 0.4524660110473633,
      "eval_runtime": 33.9797,
      "eval_samples_per_second": 279.843,
      "eval_steps_per_second": 1.118,
      "step": 2000
    },
    {
      "epoch": 1.0876811594202898,
      "grad_norm": 6.607709884643555,
      "learning_rate": 4.530211480362538e-05,
      "loss": 1.4282,
      "step": 3000
    },
    {
      "epoch": 1.0876811594202898,
      "eval_cosine_accuracy": 0.9617204666137695,
      "eval_loss": 0.4302012026309967,
      "eval_runtime": 34.5566,
      "eval_samples_per_second": 275.172,
      "eval_steps_per_second": 1.1,
      "step": 3000
    },
    {
      "epoch": 1.45,
      "grad_norm": 5.233509540557861,
      "learning_rate": 4.739764314851186e-05,
      "loss": 1.556,
      "step": 4000
    },
    {
      "epoch": 1.45,
      "eval_cosine_accuracy": 0.9664528369903564,
      "eval_loss": 0.4086464047431946,
      "eval_runtime": 33.7595,
      "eval_samples_per_second": 281.669,
      "eval_steps_per_second": 1.126,
      "step": 4000
    },
    {
      "epoch": 1.8123188405797102,
      "grad_norm": 5.165083408355713,
      "learning_rate": 4.362441456413356e-05,
      "loss": 1.3647,
      "step": 5000
    },
    {
      "epoch": 1.8123188405797102,
      "eval_cosine_accuracy": 0.9669786691665649,
      "eval_loss": 0.4053584337234497,
      "eval_runtime": 34.8117,
      "eval_samples_per_second": 273.155,
      "eval_steps_per_second": 1.092,
      "step": 5000
    },
    {
      "epoch": 2.17463768115942,
      "grad_norm": 4.441340446472168,
      "learning_rate": 3.985118597975525e-05,
      "loss": 1.2206,
      "step": 6000
    },
    {
      "epoch": 2.17463768115942,
      "eval_cosine_accuracy": 0.9669786691665649,
      "eval_loss": 0.4014289677143097,
      "eval_runtime": 34.453,
      "eval_samples_per_second": 275.999,
      "eval_steps_per_second": 1.103,
      "step": 6000
    },
    {
      "epoch": 2.5369565217391306,
      "grad_norm": 4.254669666290283,
      "learning_rate": 3.607418038978698e-05,
      "loss": 1.1389,
      "step": 7000
    },
    {
      "epoch": 2.5369565217391306,
      "eval_cosine_accuracy": 0.9693974256515503,
      "eval_loss": 0.3999493420124054,
      "eval_runtime": 33.7947,
      "eval_samples_per_second": 281.376,
      "eval_steps_per_second": 1.124,
      "step": 7000
    },
    {
      "epoch": 2.8992753623188405,
      "grad_norm": 4.167431354522705,
      "learning_rate": 3.229717479981871e-05,
      "loss": 1.1119,
      "step": 8000
    },
    {
      "epoch": 2.8992753623188405,
      "eval_cosine_accuracy": 0.9703438878059387,
      "eval_loss": 0.3831478953361511,
      "eval_runtime": 34.0399,
      "eval_samples_per_second": 279.348,
      "eval_steps_per_second": 1.116,
      "step": 8000
    },
    {
      "epoch": 3.261594202898551,
      "grad_norm": 5.115843296051025,
      "learning_rate": 2.85239462154404e-05,
      "loss": 1.0188,
      "step": 9000
    },
    {
      "epoch": 3.261594202898551,
      "eval_cosine_accuracy": 0.968766450881958,
      "eval_loss": 0.3900427520275116,
      "eval_runtime": 33.991,
      "eval_samples_per_second": 279.751,
      "eval_steps_per_second": 1.118,
      "step": 9000
    },
    {
      "epoch": 3.623913043478261,
      "grad_norm": 5.5258989334106445,
      "learning_rate": 2.4746940625472125e-05,
      "loss": 1.0139,
      "step": 10000
    },
    {
      "epoch": 3.623913043478261,
      "eval_cosine_accuracy": 0.970974862575531,
      "eval_loss": 0.3912922441959381,
      "eval_runtime": 34.0091,
      "eval_samples_per_second": 279.601,
      "eval_steps_per_second": 1.117,
      "step": 10000
    },
    {
      "epoch": 3.986231884057971,
      "grad_norm": 5.09008264541626,
      "learning_rate": 2.0969935035503853e-05,
      "loss": 0.9604,
      "step": 11000
    },
    {
      "epoch": 3.986231884057971,
      "eval_cosine_accuracy": 0.9716058373451233,
      "eval_loss": 0.3870432376861572,
      "eval_runtime": 35.0845,
      "eval_samples_per_second": 271.031,
      "eval_steps_per_second": 1.083,
      "step": 11000
    },
    {
      "epoch": 4.348550724637681,
      "grad_norm": 3.954172134399414,
      "learning_rate": 1.719292944553558e-05,
      "loss": 0.9293,
      "step": 12000
    },
    {
      "epoch": 4.348550724637681,
      "eval_cosine_accuracy": 0.9701335430145264,
      "eval_loss": 0.3847765326499939,
      "eval_runtime": 33.706,
      "eval_samples_per_second": 282.116,
      "eval_steps_per_second": 1.127,
      "step": 12000
    },
    {
      "epoch": 4.710869565217392,
      "grad_norm": 4.742279529571533,
      "learning_rate": 1.3419700861157275e-05,
      "loss": 0.9132,
      "step": 13000
    },
    {
      "epoch": 4.710869565217392,
      "eval_cosine_accuracy": 0.9704490303993225,
      "eval_loss": 0.3883793354034424,
      "eval_runtime": 33.8101,
      "eval_samples_per_second": 281.248,
      "eval_steps_per_second": 1.124,
      "step": 13000
    },
    {
      "epoch": 5.073188405797102,
      "grad_norm": 4.421851634979248,
      "learning_rate": 9.642695271189002e-06,
      "loss": 0.8993,
      "step": 14000
    },
    {
      "epoch": 5.073188405797102,
      "eval_cosine_accuracy": 0.9708697199821472,
      "eval_loss": 0.39104098081588745,
      "eval_runtime": 33.8951,
      "eval_samples_per_second": 280.542,
      "eval_steps_per_second": 1.121,
      "step": 14000
    },
    {
      "epoch": 5.435507246376812,
      "grad_norm": 4.973081588745117,
      "learning_rate": 5.869466686810697e-06,
      "loss": 0.8662,
      "step": 15000
    },
    {
      "epoch": 5.435507246376812,
      "eval_cosine_accuracy": 0.9710800051689148,
      "eval_loss": 0.38987207412719727,
      "eval_runtime": 34.7686,
      "eval_samples_per_second": 273.494,
      "eval_steps_per_second": 1.093,
      "step": 15000
    },
    {
      "epoch": 5.797826086956522,
      "grad_norm": 5.636115550994873,
      "learning_rate": 2.0924610968424236e-06,
      "loss": 0.8616,
      "step": 16000
    },
    {
      "epoch": 5.797826086956522,
      "eval_cosine_accuracy": 0.9711852073669434,
      "eval_loss": 0.39057692885398865,
      "eval_runtime": 34.4768,
      "eval_samples_per_second": 275.809,
      "eval_steps_per_second": 1.102,
      "step": 16000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 16548,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 6,
  "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
}