{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "global_step": 6960,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.57,
      "eval_cer": 0.6760756909547738,
      "eval_loss": 3.268118381500244,
      "eval_runtime": 47.2428,
      "eval_samples_per_second": 9.673,
      "eval_steps_per_second": 1.228,
      "step": 400
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000994,
      "loss": 7.285,
      "step": 500
    },
    {
      "epoch": 1.15,
      "eval_cer": 0.41700690954773867,
      "eval_loss": 1.5312176942825317,
      "eval_runtime": 44.9156,
      "eval_samples_per_second": 10.175,
      "eval_steps_per_second": 1.291,
      "step": 800
    },
    {
      "epoch": 1.44,
      "learning_rate": 0.0009854664321151051,
      "loss": 1.259,
      "step": 1000
    },
    {
      "epoch": 1.72,
      "eval_cer": 0.3845791457286432,
      "eval_loss": 1.3459197282791138,
      "eval_runtime": 44.6109,
      "eval_samples_per_second": 10.244,
      "eval_steps_per_second": 1.3,
      "step": 1200
    },
    {
      "epoch": 2.16,
      "learning_rate": 0.0009423711049473654,
      "loss": 0.9108,
      "step": 1500
    },
    {
      "epoch": 2.3,
      "eval_cer": 0.3239243090452261,
      "eval_loss": 1.1356585025787354,
      "eval_runtime": 45.0375,
      "eval_samples_per_second": 10.147,
      "eval_steps_per_second": 1.288,
      "step": 1600
    },
    {
      "epoch": 2.87,
      "learning_rate": 0.0008732489887455042,
      "loss": 0.7227,
      "step": 2000
    },
    {
      "epoch": 2.87,
      "eval_cer": 0.30559045226130654,
      "eval_loss": 1.0570871829986572,
      "eval_runtime": 44.5613,
      "eval_samples_per_second": 10.256,
      "eval_steps_per_second": 1.302,
      "step": 2000
    },
    {
      "epoch": 3.45,
      "eval_cer": 0.2828596105527638,
      "eval_loss": 1.0001708269119263,
      "eval_runtime": 44.9933,
      "eval_samples_per_second": 10.157,
      "eval_steps_per_second": 1.289,
      "step": 2400
    },
    {
      "epoch": 3.59,
      "learning_rate": 0.0007821668652587319,
      "loss": 0.5689,
      "step": 2500
    },
    {
      "epoch": 4.02,
      "eval_cer": 0.25529993718592964,
      "eval_loss": 0.8773429989814758,
      "eval_runtime": 44.5805,
      "eval_samples_per_second": 10.251,
      "eval_steps_per_second": 1.301,
      "step": 2800
    },
    {
      "epoch": 4.31,
      "learning_rate": 0.0006744835275597833,
      "loss": 0.4676,
      "step": 3000
    },
    {
      "epoch": 4.6,
      "eval_cer": 0.2461526381909548,
      "eval_loss": 0.8633947968482971,
      "eval_runtime": 45.9822,
      "eval_samples_per_second": 9.939,
      "eval_steps_per_second": 1.261,
      "step": 3200
    },
    {
      "epoch": 5.03,
      "learning_rate": 0.0005565344968466675,
      "loss": 0.3805,
      "step": 3500
    },
    {
      "epoch": 5.17,
      "eval_cer": 0.23233354271356785,
      "eval_loss": 0.8504095077514648,
      "eval_runtime": 44.4337,
      "eval_samples_per_second": 10.285,
      "eval_steps_per_second": 1.305,
      "step": 3600
    },
    {
      "epoch": 5.75,
      "learning_rate": 0.0004352592736989133,
      "loss": 0.2548,
      "step": 4000
    },
    {
      "epoch": 5.75,
      "eval_cer": 0.22597361809045227,
      "eval_loss": 0.8479630351066589,
      "eval_runtime": 45.0228,
      "eval_samples_per_second": 10.15,
      "eval_steps_per_second": 1.288,
      "step": 4000
    },
    {
      "epoch": 6.32,
      "eval_cer": 0.22314698492462312,
      "eval_loss": 0.8549605011940002,
      "eval_runtime": 45.2463,
      "eval_samples_per_second": 10.1,
      "eval_steps_per_second": 1.282,
      "step": 4400
    },
    {
      "epoch": 6.47,
      "learning_rate": 0.000317793054363817,
      "loss": 0.189,
      "step": 4500
    },
    {
      "epoch": 6.9,
      "eval_cer": 0.215884108040201,
      "eval_loss": 0.8586787581443787,
      "eval_runtime": 44.7999,
      "eval_samples_per_second": 10.201,
      "eval_steps_per_second": 1.295,
      "step": 4800
    },
    {
      "epoch": 7.18,
      "learning_rate": 0.00021104693333891388,
      "loss": 0.1336,
      "step": 5000
    },
    {
      "epoch": 7.47,
      "eval_cer": 0.2100738065326633,
      "eval_loss": 0.9012072086334229,
      "eval_runtime": 44.2547,
      "eval_samples_per_second": 10.327,
      "eval_steps_per_second": 1.311,
      "step": 5200
    },
    {
      "epoch": 7.9,
      "learning_rate": 0.0001213012909226786,
      "loss": 0.0827,
      "step": 5500
    },
    {
      "epoch": 8.05,
      "eval_cer": 0.21003454773869346,
      "eval_loss": 0.9302372932434082,
      "eval_runtime": 45.3822,
      "eval_samples_per_second": 10.07,
      "eval_steps_per_second": 1.278,
      "step": 5600
    },
    {
      "epoch": 8.62,
      "learning_rate": 5.3836288671270994e-05,
      "loss": 0.0506,
      "step": 6000
    },
    {
      "epoch": 8.62,
      "eval_cer": 0.20634422110552764,
      "eval_loss": 0.9622207283973694,
      "eval_runtime": 45.5783,
      "eval_samples_per_second": 10.027,
      "eval_steps_per_second": 1.273,
      "step": 6000
    },
    {
      "epoch": 9.2,
      "eval_cer": 0.2062264447236181,
      "eval_loss": 0.9826124310493469,
      "eval_runtime": 44.4354,
      "eval_samples_per_second": 10.285,
      "eval_steps_per_second": 1.305,
      "step": 6400
    },
    {
      "epoch": 9.34,
      "learning_rate": 1.2621212465134868e-05,
      "loss": 0.0389,
      "step": 6500
    },
    {
      "epoch": 9.77,
      "eval_cer": 0.20673680904522612,
      "eval_loss": 0.9933438301086426,
      "eval_runtime": 45.2582,
      "eval_samples_per_second": 10.098,
      "eval_steps_per_second": 1.282,
      "step": 6800
    },
    {
      "epoch": 10.0,
      "step": 6960,
      "total_flos": 4.4010774407784806e+18,
      "train_loss": 0.888924695431501,
      "train_runtime": 24300.3525,
      "train_samples_per_second": 9.162,
      "train_steps_per_second": 0.286
    }
  ],
  "max_steps": 6960,
  "num_train_epochs": 10,
  "total_flos": 4.4010774407784806e+18,
  "trial_name": null,
  "trial_params": null
}