{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.999443888332777,
  "global_step": 22475,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.11,
      "learning_rate": 3e-05,
      "loss": 0.1124,
      "step": 500
    },
    {
      "epoch": 0.22,
      "learning_rate": 2.9317406143344712e-05,
      "loss": 0.1142,
      "step": 1000
    },
    {
      "epoch": 0.22,
      "eval_cer": 0.012641370841580549,
      "eval_loss": 0.04826715216040611,
      "eval_runtime": 1346.7575,
      "eval_samples_per_second": 11.058,
      "eval_steps_per_second": 1.383,
      "eval_wer": 0.07066222104846163,
      "step": 1000
    },
    {
      "epoch": 0.33,
      "learning_rate": 2.863481228668942e-05,
      "loss": 0.113,
      "step": 1500
    },
    {
      "epoch": 0.44,
      "learning_rate": 2.7953583617747444e-05,
      "loss": 0.1049,
      "step": 2000
    },
    {
      "epoch": 0.44,
      "eval_cer": 0.01225587771877786,
      "eval_loss": 0.04738128185272217,
      "eval_runtime": 1341.1997,
      "eval_samples_per_second": 11.103,
      "eval_steps_per_second": 1.388,
      "eval_wer": 0.06750957348474845,
      "step": 2000
    },
    {
      "epoch": 0.56,
      "learning_rate": 2.727098976109215e-05,
      "loss": 0.0991,
      "step": 2500
    },
    {
      "epoch": 0.67,
      "learning_rate": 2.6588395904436862e-05,
      "loss": 0.0982,
      "step": 3000
    },
    {
      "epoch": 0.67,
      "eval_cer": 0.011982528049881409,
      "eval_loss": 0.04708929359912872,
      "eval_runtime": 1340.0171,
      "eval_samples_per_second": 11.113,
      "eval_steps_per_second": 1.39,
      "eval_wer": 0.0663871649280338,
      "step": 3000
    },
    {
      "epoch": 0.78,
      "learning_rate": 2.5905802047781573e-05,
      "loss": 0.0996,
      "step": 3500
    },
    {
      "epoch": 0.89,
      "learning_rate": 2.522457337883959e-05,
      "loss": 0.092,
      "step": 4000
    },
    {
      "epoch": 0.89,
      "eval_cer": 0.01171338376050644,
      "eval_loss": 0.04587830230593681,
      "eval_runtime": 1347.3232,
      "eval_samples_per_second": 11.053,
      "eval_steps_per_second": 1.382,
      "eval_wer": 0.0639855407368282,
      "step": 4000
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.45419795221843e-05,
      "loss": 0.0896,
      "step": 4500
    },
    {
      "epoch": 1.11,
      "learning_rate": 2.385938566552901e-05,
      "loss": 0.0847,
      "step": 5000
    },
    {
      "epoch": 1.11,
      "eval_cer": 0.011517132716170525,
      "eval_loss": 0.04586043953895569,
      "eval_runtime": 1342.7891,
      "eval_samples_per_second": 11.09,
      "eval_steps_per_second": 1.387,
      "eval_wer": 0.06308596329063779,
      "step": 5000
    },
    {
      "epoch": 1.22,
      "learning_rate": 2.317679180887372e-05,
      "loss": 0.0838,
      "step": 5500
    },
    {
      "epoch": 1.33,
      "learning_rate": 2.249419795221843e-05,
      "loss": 0.0837,
      "step": 6000
    },
    {
      "epoch": 1.33,
      "eval_cer": 0.011336301396746719,
      "eval_loss": 0.04530934989452362,
      "eval_runtime": 1343.699,
      "eval_samples_per_second": 11.083,
      "eval_steps_per_second": 1.386,
      "eval_wer": 0.06242572296315859,
      "step": 6000
    },
    {
      "epoch": 1.45,
      "learning_rate": 2.181160409556314e-05,
      "loss": 0.0822,
      "step": 6500
    },
    {
      "epoch": 1.56,
      "learning_rate": 2.1130375426621163e-05,
      "loss": 0.0803,
      "step": 7000
    },
    {
      "epoch": 1.56,
      "eval_cer": 0.010879316822078803,
      "eval_loss": 0.04430945962667465,
      "eval_runtime": 1343.5225,
      "eval_samples_per_second": 11.084,
      "eval_steps_per_second": 1.386,
      "eval_wer": 0.0597682556450548,
      "step": 7000
    },
    {
      "epoch": 1.67,
      "learning_rate": 2.044778156996587e-05,
      "loss": 0.0811,
      "step": 7500
    },
    {
      "epoch": 1.78,
      "learning_rate": 1.976518771331058e-05,
      "loss": 0.0826,
      "step": 8000
    },
    {
      "epoch": 1.78,
      "eval_cer": 0.011001272828201836,
      "eval_loss": 0.044084154069423676,
      "eval_runtime": 1339.8673,
      "eval_samples_per_second": 11.115,
      "eval_steps_per_second": 1.39,
      "eval_wer": 0.06039548395616004,
      "step": 8000
    },
    {
      "epoch": 1.89,
      "learning_rate": 1.9082593856655292e-05,
      "loss": 0.0795,
      "step": 8500
    },
    {
      "epoch": 2.0,
      "learning_rate": 1.84e-05,
      "loss": 0.0809,
      "step": 9000
    },
    {
      "epoch": 2.0,
      "eval_cer": 0.01099146027598504,
      "eval_loss": 0.04365098848938942,
      "eval_runtime": 1342.6486,
      "eval_samples_per_second": 11.092,
      "eval_steps_per_second": 1.387,
      "eval_wer": 0.06050277300937541,
      "step": 9000
    },
    {
      "epoch": 2.11,
      "learning_rate": 1.771740614334471e-05,
      "loss": 0.0726,
      "step": 9500
    },
    {
      "epoch": 2.22,
      "learning_rate": 1.7034812286689422e-05,
      "loss": 0.0728,
      "step": 10000
    },
    {
      "epoch": 2.22,
      "eval_cer": 0.010905950892381534,
      "eval_loss": 0.045100387185811996,
      "eval_runtime": 1342.5545,
      "eval_samples_per_second": 11.092,
      "eval_steps_per_second": 1.387,
      "eval_wer": 0.05974349663277433,
      "step": 10000
    },
    {
      "epoch": 2.34,
      "learning_rate": 1.6352218430034133e-05,
      "loss": 0.072,
      "step": 10500
    },
    {
      "epoch": 2.45,
      "learning_rate": 1.567098976109215e-05,
      "loss": 0.0707,
      "step": 11000
    },
    {
      "epoch": 2.45,
      "eval_cer": 0.010758762609129599,
      "eval_loss": 0.04439757764339447,
      "eval_runtime": 1341.1302,
      "eval_samples_per_second": 11.104,
      "eval_steps_per_second": 1.388,
      "eval_wer": 0.05909976231348211,
      "step": 11000
    },
    {
      "epoch": 2.56,
      "learning_rate": 1.4988395904436861e-05,
      "loss": 0.072,
      "step": 11500
    },
    {
      "epoch": 2.67,
      "learning_rate": 1.430580204778157e-05,
      "loss": 0.0698,
      "step": 12000
    },
    {
      "epoch": 2.67,
      "eval_cer": 0.010531672114969469,
      "eval_loss": 0.04415421932935715,
      "eval_runtime": 1339.1928,
      "eval_samples_per_second": 11.12,
      "eval_steps_per_second": 1.39,
      "eval_wer": 0.05758120956027994,
      "step": 12000
    },
    {
      "epoch": 2.78,
      "learning_rate": 1.3624573378839592e-05,
      "loss": 0.0958,
      "step": 12500
    },
    {
      "epoch": 2.89,
      "learning_rate": 1.29419795221843e-05,
      "loss": 0.0981,
      "step": 13000
    },
    {
      "epoch": 2.89,
      "eval_cer": 0.010388689211239017,
      "eval_loss": 0.04108743742108345,
      "eval_runtime": 1339.0236,
      "eval_samples_per_second": 11.122,
      "eval_steps_per_second": 1.391,
      "eval_wer": 0.05717681235969893,
      "step": 13000
    },
    {
      "epoch": 3.0,
      "learning_rate": 1.225938566552901e-05,
      "loss": 0.0991,
      "step": 13500
    },
    {
      "epoch": 3.11,
      "learning_rate": 1.1576791808873721e-05,
      "loss": 0.0928,
      "step": 14000
    },
    {
      "epoch": 3.11,
      "eval_cer": 0.010234491962117941,
      "eval_loss": 0.04134270176291466,
      "eval_runtime": 1343.5325,
      "eval_samples_per_second": 11.084,
      "eval_steps_per_second": 1.386,
      "eval_wer": 0.056062656807077775,
      "step": 14000
    },
    {
      "epoch": 3.23,
      "learning_rate": 1.089556313993174e-05,
      "loss": 0.0919,
      "step": 14500
    },
    {
      "epoch": 3.34,
      "learning_rate": 1.0212969283276451e-05,
      "loss": 0.0927,
      "step": 15000
    },
    {
      "epoch": 3.34,
      "eval_cer": 0.010227482996248802,
      "eval_loss": 0.041016656905412674,
      "eval_runtime": 1340.2656,
      "eval_samples_per_second": 11.111,
      "eval_steps_per_second": 1.389,
      "eval_wer": 0.0564588010035653,
      "step": 15000
    },
    {
      "epoch": 3.45,
      "learning_rate": 9.53037542662116e-06,
      "loss": 0.0896,
      "step": 15500
    },
    {
      "epoch": 3.56,
      "learning_rate": 8.84778156996587e-06,
      "loss": 0.0886,
      "step": 16000
    },
    {
      "epoch": 3.56,
      "eval_cer": 0.01016019692390506,
      "eval_loss": 0.040184538811445236,
      "eval_runtime": 1343.3925,
      "eval_samples_per_second": 11.085,
      "eval_steps_per_second": 1.386,
      "eval_wer": 0.055798560676086094,
      "step": 16000
    },
    {
      "epoch": 3.67,
      "learning_rate": 8.16518771331058e-06,
      "loss": 0.0872,
      "step": 16500
    },
    {
      "epoch": 3.78,
      "learning_rate": 7.482593856655291e-06,
      "loss": 0.091,
      "step": 17000
    },
    {
      "epoch": 3.78,
      "eval_cer": 0.010059267815389445,
      "eval_loss": 0.03996647894382477,
      "eval_runtime": 1343.622,
      "eval_samples_per_second": 11.083,
      "eval_steps_per_second": 1.386,
      "eval_wer": 0.05525386240591575,
      "step": 17000
    },
    {
      "epoch": 3.89,
      "learning_rate": 6.8e-06,
      "loss": 0.0882,
      "step": 17500
    },
    {
      "epoch": 4.0,
      "learning_rate": 6.11740614334471e-06,
      "loss": 0.0888,
      "step": 18000
    },
    {
      "epoch": 4.0,
      "eval_cer": 0.009962544086395317,
      "eval_loss": 0.039790086448192596,
      "eval_runtime": 1340.4959,
      "eval_samples_per_second": 11.109,
      "eval_steps_per_second": 1.389,
      "eval_wer": 0.05461838109071702,
      "step": 18000
    },
    {
      "epoch": 4.12,
      "learning_rate": 5.43617747440273e-06,
      "loss": 0.0872,
      "step": 18500
    },
    {
      "epoch": 4.23,
      "learning_rate": 4.75358361774744e-06,
      "loss": 0.0885,
      "step": 19000
    },
    {
      "epoch": 4.23,
      "eval_cer": 0.009899463393573059,
      "eval_loss": 0.039466630667448044,
      "eval_runtime": 1341.4804,
      "eval_samples_per_second": 11.101,
      "eval_steps_per_second": 1.388,
      "eval_wer": 0.05423048989832299,
      "step": 19000
    },
    {
      "epoch": 4.34,
      "learning_rate": 4.0709897610921505e-06,
      "loss": 0.0885,
      "step": 19500
    },
    {
      "epoch": 4.45,
      "learning_rate": 3.3897610921501705e-06,
      "loss": 0.0869,
      "step": 20000
    },
    {
      "epoch": 4.45,
      "eval_cer": 0.009868623943748843,
      "eval_loss": 0.0393984392285347,
      "eval_runtime": 1341.5672,
      "eval_samples_per_second": 11.1,
      "eval_steps_per_second": 1.388,
      "eval_wer": 0.05402416479598574,
      "step": 20000
    },
    {
      "epoch": 4.56,
      "learning_rate": 2.708532423208191e-06,
      "loss": 0.0845,
      "step": 20500
    },
    {
      "epoch": 4.67,
      "learning_rate": 2.025938566552901e-06,
      "loss": 0.0844,
      "step": 21000
    },
    {
      "epoch": 4.67,
      "eval_cer": 0.009830775528055488,
      "eval_loss": 0.03927738592028618,
      "eval_runtime": 1339.2196,
      "eval_samples_per_second": 11.12,
      "eval_steps_per_second": 1.39,
      "eval_wer": 0.05393338175095735,
      "step": 21000
    },
    {
      "epoch": 4.78,
      "learning_rate": 1.3433447098976108e-06,
      "loss": 0.0866,
      "step": 21500
    },
    {
      "epoch": 4.89,
      "learning_rate": 6.607508532423209e-07,
      "loss": 0.0882,
      "step": 22000
    },
    {
      "epoch": 4.89,
      "eval_cer": 0.009790123526014478,
      "eval_loss": 0.039073262363672256,
      "eval_runtime": 1341.9861,
      "eval_samples_per_second": 11.097,
      "eval_steps_per_second": 1.387,
      "eval_wer": 0.053735309652713587,
      "step": 22000
    },
    {
      "epoch": 5.0,
      "step": 22475,
      "total_flos": 1.3602795268583408e+20,
      "train_loss": 0.08836701542702612,
      "train_runtime": 131691.1828,
      "train_samples_per_second": 5.462,
      "train_steps_per_second": 0.171
    }
  ],
  "max_steps": 22475,
  "num_train_epochs": 5,
  "total_flos": 1.3602795268583408e+20,
  "trial_name": null,
  "trial_params": null
}