{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 7299,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.20550760378133992,
      "grad_norm": 0.022213317453861237,
      "learning_rate": 1.8629949308124403e-05,
      "loss": 0.1893,
      "step": 500
    },
    {
      "epoch": 0.20550760378133992,
      "eval_accuracy": 0.9274409044193217,
      "eval_loss": 0.384206622838974,
      "eval_runtime": 66.4904,
      "eval_samples_per_second": 73.168,
      "eval_steps_per_second": 9.159,
      "step": 500
    },
    {
      "epoch": 0.41101520756267984,
      "grad_norm": 0.016410009935498238,
      "learning_rate": 1.7259898616248804e-05,
      "loss": 0.0675,
      "step": 1000
    },
    {
      "epoch": 0.41101520756267984,
      "eval_accuracy": 0.994655704008222,
      "eval_loss": 0.028500495478510857,
      "eval_runtime": 66.4366,
      "eval_samples_per_second": 73.228,
      "eval_steps_per_second": 9.167,
      "step": 1000
    },
    {
      "epoch": 0.6165228113440198,
      "grad_norm": 0.013172637671232224,
      "learning_rate": 1.58898479243732e-05,
      "loss": 0.0519,
      "step": 1500
    },
    {
      "epoch": 0.6165228113440198,
      "eval_accuracy": 0.9852004110996917,
      "eval_loss": 0.06652244180440903,
      "eval_runtime": 37.1596,
      "eval_samples_per_second": 130.922,
      "eval_steps_per_second": 16.389,
      "step": 1500
    },
    {
      "epoch": 0.8220304151253597,
      "grad_norm": 93.64683532714844,
      "learning_rate": 1.4519797232497603e-05,
      "loss": 0.0287,
      "step": 2000
    },
    {
      "epoch": 0.8220304151253597,
      "eval_accuracy": 0.9891058581706064,
      "eval_loss": 0.05017192289233208,
      "eval_runtime": 31.6255,
      "eval_samples_per_second": 153.832,
      "eval_steps_per_second": 19.257,
      "step": 2000
    },
    {
      "epoch": 1.0275380189066996,
      "grad_norm": 0.004394204821437597,
      "learning_rate": 1.3149746540622004e-05,
      "loss": 0.0334,
      "step": 2500
    },
    {
      "epoch": 1.0275380189066996,
      "eval_accuracy": 0.9905447070914697,
      "eval_loss": 0.04894111305475235,
      "eval_runtime": 47.4094,
      "eval_samples_per_second": 102.617,
      "eval_steps_per_second": 12.846,
      "step": 2500
    },
    {
      "epoch": 1.2330456226880395,
      "grad_norm": 0.0032397848553955555,
      "learning_rate": 1.1779695848746405e-05,
      "loss": 0.0218,
      "step": 3000
    },
    {
      "epoch": 1.2330456226880395,
      "eval_accuracy": 0.9819116135662899,
      "eval_loss": 0.10269968956708908,
      "eval_runtime": 45.531,
      "eval_samples_per_second": 106.85,
      "eval_steps_per_second": 13.375,
      "step": 3000
    },
    {
      "epoch": 1.4385532264693794,
      "grad_norm": 0.0003408812917768955,
      "learning_rate": 1.0409645156870804e-05,
      "loss": 0.0102,
      "step": 3500
    },
    {
      "epoch": 1.4385532264693794,
      "eval_accuracy": 0.9944501541623844,
      "eval_loss": 0.04378344491124153,
      "eval_runtime": 45.7881,
      "eval_samples_per_second": 106.25,
      "eval_steps_per_second": 13.3,
      "step": 3500
    },
    {
      "epoch": 1.6440608302507194,
      "grad_norm": 0.0026256833225488663,
      "learning_rate": 9.039594464995205e-06,
      "loss": 0.0038,
      "step": 4000
    },
    {
      "epoch": 1.6440608302507194,
      "eval_accuracy": 0.9704008221993834,
      "eval_loss": 0.19674982130527496,
      "eval_runtime": 45.5104,
      "eval_samples_per_second": 106.899,
      "eval_steps_per_second": 13.382,
      "step": 4000
    },
    {
      "epoch": 1.8495684340320593,
      "grad_norm": 0.0002484402502886951,
      "learning_rate": 7.669543773119606e-06,
      "loss": 0.012,
      "step": 4500
    },
    {
      "epoch": 1.8495684340320593,
      "eval_accuracy": 0.9852004110996917,
      "eval_loss": 0.07537884265184402,
      "eval_runtime": 45.1736,
      "eval_samples_per_second": 107.696,
      "eval_steps_per_second": 13.481,
      "step": 4500
    },
    {
      "epoch": 2.055076037813399,
      "grad_norm": 0.0006908943178132176,
      "learning_rate": 6.299493081244007e-06,
      "loss": 0.0061,
      "step": 5000
    },
    {
      "epoch": 2.055076037813399,
      "eval_accuracy": 0.9730729701952724,
      "eval_loss": 0.166794553399086,
      "eval_runtime": 45.1304,
      "eval_samples_per_second": 107.799,
      "eval_steps_per_second": 13.494,
      "step": 5000
    },
    {
      "epoch": 2.260583641594739,
      "grad_norm": 0.00021314685000106692,
      "learning_rate": 4.929442389368407e-06,
      "loss": 0.003,
      "step": 5500
    },
    {
      "epoch": 2.260583641594739,
      "eval_accuracy": 0.9954779033915725,
      "eval_loss": 0.03057803027331829,
      "eval_runtime": 45.2961,
      "eval_samples_per_second": 107.404,
      "eval_steps_per_second": 13.445,
      "step": 5500
    },
    {
      "epoch": 2.466091245376079,
      "grad_norm": 0.00010584539995761588,
      "learning_rate": 3.5593916974928076e-06,
      "loss": 0.0,
      "step": 6000
    },
    {
      "epoch": 2.466091245376079,
      "eval_accuracy": 0.9926002055498458,
      "eval_loss": 0.05110383406281471,
      "eval_runtime": 45.54,
      "eval_samples_per_second": 106.829,
      "eval_steps_per_second": 13.373,
      "step": 6000
    },
    {
      "epoch": 2.671598849157419,
      "grad_norm": 9.755617793416604e-05,
      "learning_rate": 2.189341005617208e-06,
      "loss": 0.0,
      "step": 6500
    },
    {
      "epoch": 2.671598849157419,
      "eval_accuracy": 0.9921891058581707,
      "eval_loss": 0.05986848846077919,
      "eval_runtime": 45.0163,
      "eval_samples_per_second": 108.072,
      "eval_steps_per_second": 13.528,
      "step": 6500
    },
    {
      "epoch": 2.877106452938759,
      "grad_norm": 0.00011223769251955673,
      "learning_rate": 8.192903137416085e-07,
      "loss": 0.0003,
      "step": 7000
    },
    {
      "epoch": 2.877106452938759,
      "eval_accuracy": 0.9868448098663926,
      "eval_loss": 0.09568490833044052,
      "eval_runtime": 45.1871,
      "eval_samples_per_second": 107.663,
      "eval_steps_per_second": 13.477,
      "step": 7000
    }
  ],
  "logging_steps": 500,
  "max_steps": 7299,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 1.195019939840466e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}