{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.0,
  "eval_steps": 2000,
  "global_step": 30786,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 6.496459429610862e-05,
      "grad_norm": 8.543986320495605,
      "learning_rate": 0.0,
      "loss": 2.8055,
      "step": 1
    },
    {
      "epoch": 0.12992918859221725,
      "grad_norm": 4.228509902954102,
      "learning_rate": 9.740449584199584e-07,
      "loss": 2.5053,
      "step": 2000
    },
    {
      "epoch": 0.12992918859221725,
      "eval_cosine_accuracy": 0.9490216970443726,
      "eval_loss": 0.9158492088317871,
      "eval_runtime": 24.8196,
      "eval_samples_per_second": 380.948,
      "eval_steps_per_second": 1.491,
      "step": 2000
    },
    {
      "epoch": 0.2598583771844345,
      "grad_norm": 4.29846715927124,
      "learning_rate": 1.948577182952183e-06,
      "loss": 2.0544,
      "step": 4000
    },
    {
      "epoch": 0.2598583771844345,
      "eval_cosine_accuracy": 0.9499735832214355,
      "eval_loss": 0.9200904369354248,
      "eval_runtime": 24.5453,
      "eval_samples_per_second": 385.206,
      "eval_steps_per_second": 1.507,
      "step": 4000
    },
    {
      "epoch": 0.38978756577665175,
      "grad_norm": 4.714716911315918,
      "learning_rate": 2.9226221413721413e-06,
      "loss": 1.7583,
      "step": 6000
    },
    {
      "epoch": 0.38978756577665175,
      "eval_cosine_accuracy": 0.9456372261047363,
      "eval_loss": 0.8998962640762329,
      "eval_runtime": 24.5547,
      "eval_samples_per_second": 385.059,
      "eval_steps_per_second": 1.507,
      "step": 6000
    },
    {
      "epoch": 0.519716754368869,
      "grad_norm": 4.514920234680176,
      "learning_rate": 3.8966670997921e-06,
      "loss": 1.5402,
      "step": 8000
    },
    {
      "epoch": 0.519716754368869,
      "eval_cosine_accuracy": 0.9466948509216309,
      "eval_loss": 0.8402906060218811,
      "eval_runtime": 24.5672,
      "eval_samples_per_second": 384.862,
      "eval_steps_per_second": 1.506,
      "step": 8000
    },
    {
      "epoch": 0.6496459429610862,
      "grad_norm": 4.10648250579834,
      "learning_rate": 4.871199324324324e-06,
      "loss": 1.3738,
      "step": 10000
    },
    {
      "epoch": 0.6496459429610862,
      "eval_cosine_accuracy": 0.947012186050415,
      "eval_loss": 0.8332843780517578,
      "eval_runtime": 24.6002,
      "eval_samples_per_second": 384.346,
      "eval_steps_per_second": 1.504,
      "step": 10000
    },
    {
      "epoch": 0.7795751315533035,
      "grad_norm": 4.327205657958984,
      "learning_rate": 5.844757016632017e-06,
      "loss": 1.241,
      "step": 12000
    },
    {
      "epoch": 0.7795751315533035,
      "eval_cosine_accuracy": 0.9478582739830017,
      "eval_loss": 0.8381510972976685,
      "eval_runtime": 24.3575,
      "eval_samples_per_second": 388.176,
      "eval_steps_per_second": 1.519,
      "step": 12000
    },
    {
      "epoch": 0.9095043201455207,
      "grad_norm": 4.739709854125977,
      "learning_rate": 6.818801975051975e-06,
      "loss": 1.1337,
      "step": 14000
    },
    {
      "epoch": 0.9095043201455207,
      "eval_cosine_accuracy": 0.9480698108673096,
      "eval_loss": 0.8515253663063049,
      "eval_runtime": 24.5042,
      "eval_samples_per_second": 385.853,
      "eval_steps_per_second": 1.51,
      "step": 14000
    },
    {
      "epoch": 1.039433508737738,
      "grad_norm": 3.9595558643341064,
      "learning_rate": 7.7933341995842e-06,
      "loss": 1.3481,
      "step": 16000
    },
    {
      "epoch": 1.039433508737738,
      "eval_cosine_accuracy": 0.9481755495071411,
      "eval_loss": 0.861483097076416,
      "eval_runtime": 24.9212,
      "eval_samples_per_second": 379.395,
      "eval_steps_per_second": 1.485,
      "step": 16000
    },
    {
      "epoch": 1.1693626973299551,
      "grad_norm": 3.8362483978271484,
      "learning_rate": 8.767379158004158e-06,
      "loss": 0.9775,
      "step": 18000
    },
    {
      "epoch": 1.1693626973299551,
      "eval_cosine_accuracy": 0.9490216970443726,
      "eval_loss": 0.866976261138916,
      "eval_runtime": 24.7567,
      "eval_samples_per_second": 381.917,
      "eval_steps_per_second": 1.495,
      "step": 18000
    },
    {
      "epoch": 1.2992918859221725,
      "grad_norm": 3.660410165786743,
      "learning_rate": 9.740936850311851e-06,
      "loss": 0.9106,
      "step": 20000
    },
    {
      "epoch": 1.2992918859221725,
      "eval_cosine_accuracy": 0.9511369466781616,
      "eval_loss": 0.8142719268798828,
      "eval_runtime": 24.6537,
      "eval_samples_per_second": 383.513,
      "eval_steps_per_second": 1.501,
      "step": 20000
    },
    {
      "epoch": 1.4292210745143896,
      "grad_norm": 3.5640735626220703,
      "learning_rate": 1.0715469074844075e-05,
      "loss": 0.8581,
      "step": 22000
    },
    {
      "epoch": 1.4292210745143896,
      "eval_cosine_accuracy": 0.9521946310997009,
      "eval_loss": 0.7982929944992065,
      "eval_runtime": 24.9083,
      "eval_samples_per_second": 379.592,
      "eval_steps_per_second": 1.485,
      "step": 22000
    },
    {
      "epoch": 1.559150263106607,
      "grad_norm": 3.663590908050537,
      "learning_rate": 1.1689514033264035e-05,
      "loss": 0.8119,
      "step": 24000
    },
    {
      "epoch": 1.559150263106607,
      "eval_cosine_accuracy": 0.9538868069648743,
      "eval_loss": 0.7904353141784668,
      "eval_runtime": 24.419,
      "eval_samples_per_second": 387.198,
      "eval_steps_per_second": 1.515,
      "step": 24000
    },
    {
      "epoch": 1.6890794516988241,
      "grad_norm": 3.101531505584717,
      "learning_rate": 1.2664046257796258e-05,
      "loss": 0.775,
      "step": 26000
    },
    {
      "epoch": 1.6890794516988241,
      "eval_cosine_accuracy": 0.9561078548431396,
      "eval_loss": 0.7433677315711975,
      "eval_runtime": 24.4716,
      "eval_samples_per_second": 386.366,
      "eval_steps_per_second": 1.512,
      "step": 26000
    },
    {
      "epoch": 1.8190086402910413,
      "grad_norm": 3.2498459815979004,
      "learning_rate": 1.3638091216216216e-05,
      "loss": 0.7376,
      "step": 28000
    },
    {
      "epoch": 1.8190086402910413,
      "eval_cosine_accuracy": 0.9565309286117554,
      "eval_loss": 0.779238760471344,
      "eval_runtime": 24.4481,
      "eval_samples_per_second": 386.737,
      "eval_steps_per_second": 1.513,
      "step": 28000
    },
    {
      "epoch": 1.9489378288832586,
      "grad_norm": 2.890700101852417,
      "learning_rate": 1.4612136174636175e-05,
      "loss": 0.7072,
      "step": 30000
    },
    {
      "epoch": 1.9489378288832586,
      "eval_cosine_accuracy": 0.9578000903129578,
      "eval_loss": 0.7882058024406433,
      "eval_runtime": 24.4341,
      "eval_samples_per_second": 386.959,
      "eval_steps_per_second": 1.514,
      "step": 30000
    }
  ],
  "logging_steps": 2000,
  "max_steps": 153930,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}