{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "eval_steps": 5000,
  "global_step": 43950,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.11376564277588168,
      "grad_norm": 21.099435806274414,
      "learning_rate": 7.984000000000001e-05,
      "loss": 4.3757,
      "step": 1000
    },
    {
      "epoch": 0.22753128555176336,
      "grad_norm": 6.149295330047607,
      "learning_rate": 0.00015984000000000001,
      "loss": 2.2482,
      "step": 2000
    },
    {
      "epoch": 0.3412969283276451,
      "grad_norm": 3.9191203117370605,
      "learning_rate": 0.00023984,
      "loss": 1.6084,
      "step": 3000
    },
    {
      "epoch": 0.4550625711035267,
      "grad_norm": 2.940685987472534,
      "learning_rate": 0.00031984,
      "loss": 1.4021,
      "step": 4000
    },
    {
      "epoch": 0.5688282138794084,
      "grad_norm": 2.688676118850708,
      "learning_rate": 0.00039984000000000005,
      "loss": 1.3104,
      "step": 5000
    },
    {
      "epoch": 0.5688282138794084,
      "eval_accuracy": 0.682576,
      "eval_loss": 1.2636802196502686,
      "eval_runtime": 30.6933,
      "eval_samples_per_second": 8145.096,
      "eval_steps_per_second": 15.932,
      "step": 5000
    },
    {
      "epoch": 0.6825938566552902,
      "grad_norm": 2.202869176864624,
      "learning_rate": 0.00047984000000000004,
      "loss": 1.2594,
      "step": 6000
    },
    {
      "epoch": 0.7963594994311718,
      "grad_norm": 1.9557334184646606,
      "learning_rate": 0.00055984,
      "loss": 1.2241,
      "step": 7000
    },
    {
      "epoch": 0.9101251422070534,
      "grad_norm": 1.6525850296020508,
      "learning_rate": 0.00063984,
      "loss": 1.197,
      "step": 8000
    },
    {
      "epoch": 1.023890784982935,
      "grad_norm": 1.422629714012146,
      "learning_rate": 0.0007197600000000001,
      "loss": 1.1731,
      "step": 9000
    },
    {
      "epoch": 1.1376564277588168,
      "grad_norm": 1.258188247680664,
      "learning_rate": 0.00079976,
      "loss": 1.1479,
      "step": 10000
    },
    {
      "epoch": 1.1376564277588168,
      "eval_accuracy": 0.709628,
      "eval_loss": 1.1420985460281372,
      "eval_runtime": 30.2541,
      "eval_samples_per_second": 8263.34,
      "eval_steps_per_second": 16.163,
      "step": 10000
    },
    {
      "epoch": 1.2514220705346986,
      "grad_norm": 1.3610972166061401,
      "learning_rate": 0.0007765066273932253,
      "loss": 1.1287,
      "step": 11000
    },
    {
      "epoch": 1.36518771331058,
      "grad_norm": 1.2322840690612793,
      "learning_rate": 0.0007529425625920472,
      "loss": 1.0996,
      "step": 12000
    },
    {
      "epoch": 1.4789533560864618,
      "grad_norm": 1.0880305767059326,
      "learning_rate": 0.000729378497790869,
      "loss": 1.0643,
      "step": 13000
    },
    {
      "epoch": 1.5927189988623436,
      "grad_norm": 1.0738807916641235,
      "learning_rate": 0.0007058379970544919,
      "loss": 1.0453,
      "step": 14000
    },
    {
      "epoch": 1.7064846416382253,
      "grad_norm": 0.9920330047607422,
      "learning_rate": 0.0006822739322533137,
      "loss": 1.0236,
      "step": 15000
    },
    {
      "epoch": 1.7064846416382253,
      "eval_accuracy": 0.740376,
      "eval_loss": 1.0128123760223389,
      "eval_runtime": 30.6186,
      "eval_samples_per_second": 8164.961,
      "eval_steps_per_second": 15.971,
      "step": 15000
    },
    {
      "epoch": 1.820250284414107,
      "grad_norm": 0.9682360291481018,
      "learning_rate": 0.0006587098674521356,
      "loss": 1.008,
      "step": 16000
    },
    {
      "epoch": 1.9340159271899886,
      "grad_norm": 0.9743279814720154,
      "learning_rate": 0.0006351458026509574,
      "loss": 0.9879,
      "step": 17000
    },
    {
      "epoch": 2.04778156996587,
      "grad_norm": 0.8997290134429932,
      "learning_rate": 0.0006116288659793814,
      "loss": 0.9578,
      "step": 18000
    },
    {
      "epoch": 2.161547212741752,
      "grad_norm": 0.8797656297683716,
      "learning_rate": 0.0005880648011782032,
      "loss": 0.9252,
      "step": 19000
    },
    {
      "epoch": 2.2753128555176336,
      "grad_norm": 0.8649430871009827,
      "learning_rate": 0.000564500736377025,
      "loss": 0.9206,
      "step": 20000
    },
    {
      "epoch": 2.2753128555176336,
      "eval_accuracy": 0.757684,
      "eval_loss": 0.9457069039344788,
      "eval_runtime": 30.4556,
      "eval_samples_per_second": 8208.682,
      "eval_steps_per_second": 16.056,
      "step": 20000
    },
    {
      "epoch": 2.3890784982935154,
      "grad_norm": 0.888418972492218,
      "learning_rate": 0.0005409366715758468,
      "loss": 0.9164,
      "step": 21000
    },
    {
      "epoch": 2.502844141069397,
      "grad_norm": 0.8337541818618774,
      "learning_rate": 0.0005173961708394698,
      "loss": 0.9029,
      "step": 22000
    },
    {
      "epoch": 2.616609783845279,
      "grad_norm": 0.8450725078582764,
      "learning_rate": 0.0004938321060382916,
      "loss": 0.9027,
      "step": 23000
    },
    {
      "epoch": 2.73037542662116,
      "grad_norm": 0.8347210884094238,
      "learning_rate": 0.00047029160530191456,
      "loss": 0.8942,
      "step": 24000
    },
    {
      "epoch": 2.8441410693970424,
      "grad_norm": 0.9702669382095337,
      "learning_rate": 0.00044672754050073635,
      "loss": 0.8878,
      "step": 25000
    },
    {
      "epoch": 2.8441410693970424,
      "eval_accuracy": 0.765228,
      "eval_loss": 0.9111341834068298,
      "eval_runtime": 30.5593,
      "eval_samples_per_second": 8180.815,
      "eval_steps_per_second": 16.002,
      "step": 25000
    },
    {
      "epoch": 2.9579067121729237,
      "grad_norm": 0.7944394946098328,
      "learning_rate": 0.00042318703976435934,
      "loss": 0.8766,
      "step": 26000
    },
    {
      "epoch": 3.0716723549488054,
      "grad_norm": 0.750736653804779,
      "learning_rate": 0.0003996465390279823,
      "loss": 0.836,
      "step": 27000
    },
    {
      "epoch": 3.185437997724687,
      "grad_norm": 0.907810628414154,
      "learning_rate": 0.0003760824742268041,
      "loss": 0.8147,
      "step": 28000
    },
    {
      "epoch": 3.299203640500569,
      "grad_norm": 0.8858300447463989,
      "learning_rate": 0.0003525184094256259,
      "loss": 0.8098,
      "step": 29000
    },
    {
      "epoch": 3.4129692832764507,
      "grad_norm": 0.8472805619239807,
      "learning_rate": 0.0003289543446244477,
      "loss": 0.8107,
      "step": 30000
    },
    {
      "epoch": 3.4129692832764507,
      "eval_accuracy": 0.7749,
      "eval_loss": 0.8753725290298462,
      "eval_runtime": 30.6567,
      "eval_samples_per_second": 8154.821,
      "eval_steps_per_second": 15.951,
      "step": 30000
    },
    {
      "epoch": 3.526734926052332,
      "grad_norm": 0.8683416843414307,
      "learning_rate": 0.0003054138438880707,
      "loss": 0.8104,
      "step": 31000
    },
    {
      "epoch": 3.640500568828214,
      "grad_norm": 0.8718369007110596,
      "learning_rate": 0.00028184977908689247,
      "loss": 0.8076,
      "step": 32000
    },
    {
      "epoch": 3.7542662116040955,
      "grad_norm": 0.7670469284057617,
      "learning_rate": 0.00025830927835051546,
      "loss": 0.7941,
      "step": 33000
    },
    {
      "epoch": 3.868031854379977,
      "grad_norm": 0.7913306951522827,
      "learning_rate": 0.00023474521354933727,
      "loss": 0.7934,
      "step": 34000
    },
    {
      "epoch": 3.981797497155859,
      "grad_norm": 0.8316972255706787,
      "learning_rate": 0.00021120471281296026,
      "loss": 0.7874,
      "step": 35000
    },
    {
      "epoch": 3.981797497155859,
      "eval_accuracy": 0.782728,
      "eval_loss": 0.8436186909675598,
      "eval_runtime": 30.5157,
      "eval_samples_per_second": 8192.491,
      "eval_steps_per_second": 16.025,
      "step": 35000
    },
    {
      "epoch": 4.09556313993174,
      "grad_norm": 0.8810792565345764,
      "learning_rate": 0.00018766421207658322,
      "loss": 0.7286,
      "step": 36000
    },
    {
      "epoch": 4.2093287827076225,
      "grad_norm": 0.8747660517692566,
      "learning_rate": 0.000164100147275405,
      "loss": 0.7156,
      "step": 37000
    },
    {
      "epoch": 4.323094425483504,
      "grad_norm": 0.8973079323768616,
      "learning_rate": 0.0001405360824742268,
      "loss": 0.7107,
      "step": 38000
    },
    {
      "epoch": 4.436860068259386,
      "grad_norm": 0.8924602270126343,
      "learning_rate": 0.00011697201767304862,
      "loss": 0.7124,
      "step": 39000
    },
    {
      "epoch": 4.550625711035267,
      "grad_norm": 0.8596675992012024,
      "learning_rate": 9.343151693667158e-05,
      "loss": 0.7064,
      "step": 40000
    },
    {
      "epoch": 4.550625711035267,
      "eval_accuracy": 0.786888,
      "eval_loss": 0.8359516263008118,
      "eval_runtime": 30.4174,
      "eval_samples_per_second": 8218.976,
      "eval_steps_per_second": 16.076,
      "step": 40000
    },
    {
      "epoch": 4.664391353811149,
      "grad_norm": 0.927306592464447,
      "learning_rate": 6.986745213549337e-05,
      "loss": 0.7033,
      "step": 41000
    },
    {
      "epoch": 4.778156996587031,
      "grad_norm": 0.9652923941612244,
      "learning_rate": 4.6326951399116346e-05,
      "loss": 0.7012,
      "step": 42000
    },
    {
      "epoch": 4.891922639362912,
      "grad_norm": 0.9106791019439697,
      "learning_rate": 2.2762886597938146e-05,
      "loss": 0.6982,
      "step": 43000
    },
    {
      "epoch": 5.0,
      "step": 43950,
      "total_flos": 5.7622608792e+17,
      "train_loss": 1.0485419180287436,
      "train_runtime": 8277.8936,
      "train_samples_per_second": 2718.083,
      "train_steps_per_second": 5.309
    }
  ],
  "logging_steps": 1000,
  "max_steps": 43950,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 5000,
  "total_flos": 5.7622608792e+17,
  "train_batch_size": 512,
  "trial_name": null,
  "trial_params": null
}