{
  "best_global_step": 90,
  "best_metric": 46.236559139784944,
  "best_model_checkpoint": "./tiny-naija-checkpoints/checkpoint-90",
  "epoch": 20.0,
  "eval_steps": 500,
  "global_step": 300,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.7017543859649122,
      "grad_norm": 79.26960754394531,
      "learning_rate": 1.8000000000000001e-06,
      "loss": 4.3206,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_loss": 3.6388587951660156,
      "eval_runtime": 14.5539,
      "eval_samples_per_second": 3.436,
      "eval_steps_per_second": 0.481,
      "eval_wer": 81.87403993855607,
      "step": 15
    },
    {
      "epoch": 1.3508771929824561,
      "grad_norm": 53.15379333496094,
      "learning_rate": 3.8000000000000005e-06,
      "loss": 3.4837,
      "step": 20
    },
    {
      "epoch": 2.0,
      "grad_norm": 58.95879364013672,
      "learning_rate": 5.8e-06,
      "loss": 2.5758,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_loss": 2.2313578128814697,
      "eval_runtime": 13.3093,
      "eval_samples_per_second": 3.757,
      "eval_steps_per_second": 0.526,
      "eval_wer": 74.03993855606758,
      "step": 30
    },
    {
      "epoch": 2.7017543859649122,
      "grad_norm": 16.26763916015625,
      "learning_rate": 7.800000000000002e-06,
      "loss": 1.629,
      "step": 40
    },
    {
      "epoch": 3.0,
      "eval_loss": 1.3885557651519775,
      "eval_runtime": 14.6694,
      "eval_samples_per_second": 3.408,
      "eval_steps_per_second": 0.477,
      "eval_wer": 61.136712749615974,
      "step": 45
    },
    {
      "epoch": 3.3508771929824563,
      "grad_norm": 12.690998077392578,
      "learning_rate": 9.800000000000001e-06,
      "loss": 1.0623,
      "step": 50
    },
    {
      "epoch": 4.0,
      "grad_norm": 34.526302337646484,
      "learning_rate": 9.640000000000001e-06,
      "loss": 0.7805,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_loss": 1.251920461654663,
      "eval_runtime": 12.2314,
      "eval_samples_per_second": 4.088,
      "eval_steps_per_second": 0.572,
      "eval_wer": 52.22734254992319,
      "step": 60
    },
    {
      "epoch": 4.701754385964913,
      "grad_norm": 8.408075332641602,
      "learning_rate": 9.240000000000001e-06,
      "loss": 0.6595,
      "step": 70
    },
    {
      "epoch": 5.0,
      "eval_loss": 1.205701231956482,
      "eval_runtime": 11.0287,
      "eval_samples_per_second": 4.534,
      "eval_steps_per_second": 0.635,
      "eval_wer": 47.15821812596006,
      "step": 75
    },
    {
      "epoch": 5.350877192982456,
      "grad_norm": 7.180197238922119,
      "learning_rate": 8.84e-06,
      "loss": 0.5661,
      "step": 80
    },
    {
      "epoch": 6.0,
      "grad_norm": 24.096446990966797,
      "learning_rate": 8.44e-06,
      "loss": 0.4548,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_loss": 1.2086139917373657,
      "eval_runtime": 11.1174,
      "eval_samples_per_second": 4.497,
      "eval_steps_per_second": 0.63,
      "eval_wer": 46.236559139784944,
      "step": 90
    },
    {
      "epoch": 6.701754385964913,
      "grad_norm": 6.298120498657227,
      "learning_rate": 8.040000000000001e-06,
      "loss": 0.3583,
      "step": 100
    },
    {
      "epoch": 7.0,
      "eval_loss": 1.2257176637649536,
      "eval_runtime": 10.9626,
      "eval_samples_per_second": 4.561,
      "eval_steps_per_second": 0.639,
      "eval_wer": 49.76958525345622,
      "step": 105
    },
    {
      "epoch": 7.350877192982456,
      "grad_norm": 5.97155237197876,
      "learning_rate": 7.640000000000001e-06,
      "loss": 0.3552,
      "step": 110
    },
    {
      "epoch": 8.0,
      "grad_norm": 35.86328125,
      "learning_rate": 7.24e-06,
      "loss": 0.2761,
      "step": 120
    },
    {
      "epoch": 8.0,
      "eval_loss": 1.240835428237915,
      "eval_runtime": 11.012,
      "eval_samples_per_second": 4.54,
      "eval_steps_per_second": 0.636,
      "eval_wer": 49.46236559139785,
      "step": 120
    },
    {
      "epoch": 8.701754385964913,
      "grad_norm": 4.273914337158203,
      "learning_rate": 6.8400000000000014e-06,
      "loss": 0.2189,
      "step": 130
    },
    {
      "epoch": 9.0,
      "eval_loss": 1.2799654006958008,
      "eval_runtime": 10.9803,
      "eval_samples_per_second": 4.554,
      "eval_steps_per_second": 0.638,
      "eval_wer": 49.76958525345622,
      "step": 135
    },
    {
      "epoch": 9.350877192982455,
      "grad_norm": 4.224545001983643,
      "learning_rate": 6.440000000000001e-06,
      "loss": 0.2044,
      "step": 140
    },
    {
      "epoch": 10.0,
      "grad_norm": 16.503740310668945,
      "learning_rate": 6.040000000000001e-06,
      "loss": 0.1708,
      "step": 150
    },
    {
      "epoch": 10.0,
      "eval_loss": 1.3069055080413818,
      "eval_runtime": 10.9717,
      "eval_samples_per_second": 4.557,
      "eval_steps_per_second": 0.638,
      "eval_wer": 48.23348694316436,
      "step": 150
    },
    {
      "epoch": 10.701754385964913,
      "grad_norm": 2.8711295127868652,
      "learning_rate": 5.64e-06,
      "loss": 0.1448,
      "step": 160
    },
    {
      "epoch": 11.0,
      "eval_loss": 1.326417326927185,
      "eval_runtime": 13.3027,
      "eval_samples_per_second": 3.759,
      "eval_steps_per_second": 0.526,
      "eval_wer": 63.594470046082954,
      "step": 165
    },
    {
      "epoch": 11.350877192982455,
      "grad_norm": 2.82039737701416,
      "learning_rate": 5.240000000000001e-06,
      "loss": 0.125,
      "step": 170
    },
    {
      "epoch": 12.0,
      "grad_norm": 19.702402114868164,
      "learning_rate": 4.84e-06,
      "loss": 0.1169,
      "step": 180
    },
    {
      "epoch": 12.0,
      "eval_loss": 1.3684873580932617,
      "eval_runtime": 13.3579,
      "eval_samples_per_second": 3.743,
      "eval_steps_per_second": 0.524,
      "eval_wer": 63.90168970814132,
      "step": 180
    },
    {
      "epoch": 12.701754385964913,
      "grad_norm": 2.337428092956543,
      "learning_rate": 4.440000000000001e-06,
      "loss": 0.0839,
      "step": 190
    },
    {
      "epoch": 13.0,
      "eval_loss": 1.3854682445526123,
      "eval_runtime": 12.277,
      "eval_samples_per_second": 4.073,
      "eval_steps_per_second": 0.57,
      "eval_wer": 63.594470046082954,
      "step": 195
    },
    {
      "epoch": 13.350877192982455,
      "grad_norm": 3.5158214569091797,
      "learning_rate": 4.04e-06,
      "loss": 0.0844,
      "step": 200
    },
    {
      "epoch": 14.0,
      "grad_norm": 2.8117311000823975,
      "learning_rate": 3.6400000000000003e-06,
      "loss": 0.0701,
      "step": 210
    },
    {
      "epoch": 14.0,
      "eval_loss": 1.4106031656265259,
      "eval_runtime": 11.1252,
      "eval_samples_per_second": 4.494,
      "eval_steps_per_second": 0.629,
      "eval_wer": 50.84485407066052,
      "step": 210
    },
    {
      "epoch": 14.701754385964913,
      "grad_norm": 1.6476364135742188,
      "learning_rate": 3.2400000000000003e-06,
      "loss": 0.06,
      "step": 220
    },
    {
      "epoch": 15.0,
      "eval_loss": 1.4286068677902222,
      "eval_runtime": 10.9778,
      "eval_samples_per_second": 4.555,
      "eval_steps_per_second": 0.638,
      "eval_wer": 48.694316436251924,
      "step": 225
    },
    {
      "epoch": 15.350877192982455,
      "grad_norm": 1.7225351333618164,
      "learning_rate": 2.84e-06,
      "loss": 0.0583,
      "step": 230
    },
    {
      "epoch": 16.0,
      "grad_norm": 24.52082633972168,
      "learning_rate": 2.4400000000000004e-06,
      "loss": 0.0796,
      "step": 240
    },
    {
      "epoch": 16.0,
      "eval_loss": 1.4395389556884766,
      "eval_runtime": 10.9741,
      "eval_samples_per_second": 4.556,
      "eval_steps_per_second": 0.638,
      "eval_wer": 48.23348694316436,
      "step": 240
    },
    {
      "epoch": 16.70175438596491,
      "grad_norm": 1.8047239780426025,
      "learning_rate": 2.04e-06,
      "loss": 0.0522,
      "step": 250
    },
    {
      "epoch": 17.0,
      "eval_loss": 1.45657479763031,
      "eval_runtime": 10.9906,
      "eval_samples_per_second": 4.549,
      "eval_steps_per_second": 0.637,
      "eval_wer": 49.0015360983103,
      "step": 255
    },
    {
      "epoch": 17.350877192982455,
      "grad_norm": 1.3042881488800049,
      "learning_rate": 1.6400000000000002e-06,
      "loss": 0.0381,
      "step": 260
    },
    {
      "epoch": 18.0,
      "grad_norm": 2.017195701599121,
      "learning_rate": 1.2400000000000002e-06,
      "loss": 0.0373,
      "step": 270
    },
    {
      "epoch": 18.0,
      "eval_loss": 1.4669005870819092,
      "eval_runtime": 10.9521,
      "eval_samples_per_second": 4.565,
      "eval_steps_per_second": 0.639,
      "eval_wer": 48.54070660522273,
      "step": 270
    },
    {
      "epoch": 18.70175438596491,
      "grad_norm": 0.8439742922782898,
      "learning_rate": 8.400000000000001e-07,
      "loss": 0.0385,
      "step": 280
    },
    {
      "epoch": 19.0,
      "eval_loss": 1.4718369245529175,
      "eval_runtime": 10.9473,
      "eval_samples_per_second": 4.567,
      "eval_steps_per_second": 0.639,
      "eval_wer": 48.8479262672811,
      "step": 285
    },
    {
      "epoch": 19.350877192982455,
      "grad_norm": 1.529391884803772,
      "learning_rate": 4.4e-07,
      "loss": 0.0306,
      "step": 290
    },
    {
      "epoch": 20.0,
      "grad_norm": 2.573256254196167,
      "learning_rate": 4e-08,
      "loss": 0.035,
      "step": 300
    },
    {
      "epoch": 20.0,
      "eval_loss": 1.4741815328598022,
      "eval_runtime": 10.9806,
      "eval_samples_per_second": 4.553,
      "eval_steps_per_second": 0.637,
      "eval_wer": 49.30875576036866,
      "step": 300
    }
  ],
  "logging_steps": 10,
  "max_steps": 300,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.2156996608e+17,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}