{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 200.3030303030303,
  "global_step": 9815,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 2.04,
      "learning_rate": 8.16743119266055e-06,
      "loss": 15.5589,
      "step": 100
    },
    {
      "epoch": 4.08,
      "learning_rate": 1.573623853211009e-05,
      "loss": 6.7324,
      "step": 200
    },
    {
      "epoch": 6.12,
      "learning_rate": 2.330504587155963e-05,
      "loss": 4.9014,
      "step": 300
    },
    {
      "epoch": 8.16,
      "learning_rate": 3.087385321100917e-05,
      "loss": 3.9471,
      "step": 400
    },
    {
      "epoch": 8.16,
      "eval_cer": 1.0,
      "eval_loss": 3.7109458446502686,
      "eval_runtime": 120.434,
      "eval_samples_per_second": 23.573,
      "eval_steps_per_second": 0.374,
      "eval_wer": 1.0,
      "step": 400
    },
    {
      "epoch": 10.2,
      "learning_rate": 3.8442660550458714e-05,
      "loss": 3.5201,
      "step": 500
    },
    {
      "epoch": 12.24,
      "learning_rate": 4.601146788990825e-05,
      "loss": 3.4075,
      "step": 600
    },
    {
      "epoch": 14.28,
      "learning_rate": 5.35802752293578e-05,
      "loss": 3.3368,
      "step": 700
    },
    {
      "epoch": 16.32,
      "learning_rate": 6.114908256880733e-05,
      "loss": 3.274,
      "step": 800
    },
    {
      "epoch": 16.32,
      "eval_cer": 0.9572537581332735,
      "eval_loss": 3.1581757068634033,
      "eval_runtime": 118.5811,
      "eval_samples_per_second": 23.941,
      "eval_steps_per_second": 0.379,
      "eval_wer": 0.9917300840188272,
      "step": 800
    },
    {
      "epoch": 18.36,
      "learning_rate": 6.871788990825688e-05,
      "loss": 2.8673,
      "step": 900
    },
    {
      "epoch": 20.4,
      "learning_rate": 7.5e-05,
      "loss": 2.0372,
      "step": 1000
    },
    {
      "epoch": 22.44,
      "learning_rate": 7.5e-05,
      "loss": 1.7262,
      "step": 1100
    },
    {
      "epoch": 24.48,
      "learning_rate": 7.5e-05,
      "loss": 1.5889,
      "step": 1200
    },
    {
      "epoch": 24.48,
      "eval_cer": 0.19903073816468478,
      "eval_loss": 0.7762717604637146,
      "eval_runtime": 118.6201,
      "eval_samples_per_second": 23.934,
      "eval_steps_per_second": 0.379,
      "eval_wer": 0.6030440329037082,
      "step": 1200
    },
    {
      "epoch": 26.53,
      "learning_rate": 7.5e-05,
      "loss": 1.5113,
      "step": 1300
    },
    {
      "epoch": 28.57,
      "learning_rate": 7.5e-05,
      "loss": 1.4529,
      "step": 1400
    },
    {
      "epoch": 30.61,
      "learning_rate": 7.5e-05,
      "loss": 1.4042,
      "step": 1500
    },
    {
      "epoch": 32.65,
      "learning_rate": 7.5e-05,
      "loss": 1.3647,
      "step": 1600
    },
    {
      "epoch": 32.65,
      "eval_cer": 0.16868745793134396,
      "eval_loss": 0.605066180229187,
      "eval_runtime": 119.2577,
      "eval_samples_per_second": 23.806,
      "eval_steps_per_second": 0.377,
      "eval_wer": 0.5134826023841992,
      "step": 1600
    },
    {
      "epoch": 34.69,
      "learning_rate": 7.5e-05,
      "loss": 1.3249,
      "step": 1700
    },
    {
      "epoch": 36.73,
      "learning_rate": 7.5e-05,
      "loss": 1.3121,
      "step": 1800
    },
    {
      "epoch": 38.77,
      "learning_rate": 7.5e-05,
      "loss": 1.2899,
      "step": 1900
    },
    {
      "epoch": 40.81,
      "learning_rate": 7.5e-05,
      "loss": 1.2532,
      "step": 2000
    },
    {
      "epoch": 40.81,
      "eval_cer": 0.15389724029616333,
      "eval_loss": 0.5423398017883301,
      "eval_runtime": 118.7297,
      "eval_samples_per_second": 23.911,
      "eval_steps_per_second": 0.379,
      "eval_wer": 0.4711652663528791,
      "step": 2000
    },
    {
      "epoch": 42.85,
      "learning_rate": 7.5e-05,
      "loss": 1.2254,
      "step": 2100
    },
    {
      "epoch": 44.89,
      "learning_rate": 7.5e-05,
      "loss": 1.2028,
      "step": 2200
    },
    {
      "epoch": 46.93,
      "learning_rate": 7.5e-05,
      "loss": 1.1903,
      "step": 2300
    },
    {
      "epoch": 48.97,
      "learning_rate": 7.5e-05,
      "loss": 1.1905,
      "step": 2400
    },
    {
      "epoch": 48.97,
      "eval_cer": 0.1490329818263406,
      "eval_loss": 0.5179790258407593,
      "eval_runtime": 119.6696,
      "eval_samples_per_second": 23.724,
      "eval_steps_per_second": 0.376,
      "eval_wer": 0.4532177891171425,
      "step": 2400
    },
    {
      "epoch": 51.02,
      "learning_rate": 7.5e-05,
      "loss": 1.1671,
      "step": 2500
    },
    {
      "epoch": 53.06,
      "learning_rate": 7.5e-05,
      "loss": 1.1381,
      "step": 2600
    },
    {
      "epoch": 55.1,
      "learning_rate": 7.5e-05,
      "loss": 1.131,
      "step": 2700
    },
    {
      "epoch": 57.14,
      "learning_rate": 7.5e-05,
      "loss": 1.1193,
      "step": 2800
    },
    {
      "epoch": 57.14,
      "eval_cer": 0.13925959165357865,
      "eval_loss": 0.4906373620033264,
      "eval_runtime": 119.1486,
      "eval_samples_per_second": 23.827,
      "eval_steps_per_second": 0.378,
      "eval_wer": 0.4248009501605595,
      "step": 2800
    },
    {
      "epoch": 59.18,
      "learning_rate": 7.5e-05,
      "loss": 1.1008,
      "step": 2900
    },
    {
      "epoch": 61.22,
      "learning_rate": 7.5e-05,
      "loss": 1.0784,
      "step": 3000
    },
    {
      "epoch": 63.26,
      "learning_rate": 7.5e-05,
      "loss": 1.0678,
      "step": 3100
    },
    {
      "epoch": 65.3,
      "learning_rate": 7.5e-05,
      "loss": 1.0584,
      "step": 3200
    },
    {
      "epoch": 65.3,
      "eval_cer": 0.13315683194974198,
      "eval_loss": 0.48543456196784973,
      "eval_runtime": 119.1946,
      "eval_samples_per_second": 23.818,
      "eval_steps_per_second": 0.378,
      "eval_wer": 0.40685347292482293,
      "step": 3200
    },
    {
      "epoch": 67.34,
      "learning_rate": 7.5e-05,
      "loss": 1.0434,
      "step": 3300
    },
    {
      "epoch": 69.38,
      "learning_rate": 7.5e-05,
      "loss": 1.0332,
      "step": 3400
    },
    {
      "epoch": 71.42,
      "learning_rate": 7.5e-05,
      "loss": 1.0198,
      "step": 3500
    },
    {
      "epoch": 73.46,
      "learning_rate": 7.5e-05,
      "loss": 1.0095,
      "step": 3600
    },
    {
      "epoch": 73.46,
      "eval_cer": 0.12868745793134395,
      "eval_loss": 0.4780372381210327,
      "eval_runtime": 118.9725,
      "eval_samples_per_second": 23.863,
      "eval_steps_per_second": 0.378,
      "eval_wer": 0.39255707561694453,
      "step": 3600
    },
    {
      "epoch": 75.51,
      "learning_rate": 7.5e-05,
      "loss": 0.9989,
      "step": 3700
    },
    {
      "epoch": 77.55,
      "learning_rate": 7.5e-05,
      "loss": 0.9924,
      "step": 3800
    },
    {
      "epoch": 79.59,
      "learning_rate": 7.5e-05,
      "loss": 0.9878,
      "step": 3900
    },
    {
      "epoch": 81.63,
      "learning_rate": 7.5e-05,
      "loss": 0.9759,
      "step": 4000
    },
    {
      "epoch": 81.63,
      "eval_cer": 0.12686560466681623,
      "eval_loss": 0.46656644344329834,
      "eval_runtime": 116.6705,
      "eval_samples_per_second": 24.333,
      "eval_steps_per_second": 0.386,
      "eval_wer": 0.39251308670215107,
      "step": 4000
    },
    {
      "epoch": 83.67,
      "learning_rate": 7.5e-05,
      "loss": 0.9659,
      "step": 4100
    },
    {
      "epoch": 85.71,
      "learning_rate": 7.5e-05,
      "loss": 0.9531,
      "step": 4200
    },
    {
      "epoch": 87.75,
      "learning_rate": 7.5e-05,
      "loss": 0.9452,
      "step": 4300
    },
    {
      "epoch": 89.79,
      "learning_rate": 7.5e-05,
      "loss": 0.9593,
      "step": 4400
    },
    {
      "epoch": 89.79,
      "eval_cer": 0.1247386134170967,
      "eval_loss": 0.4808085262775421,
      "eval_runtime": 119.5296,
      "eval_samples_per_second": 23.751,
      "eval_steps_per_second": 0.376,
      "eval_wer": 0.38296749219196763,
      "step": 4400
    },
    {
      "epoch": 91.83,
      "learning_rate": 7.5e-05,
      "loss": 0.928,
      "step": 4500
    },
    {
      "epoch": 93.87,
      "learning_rate": 7.5e-05,
      "loss": 0.9251,
      "step": 4600
    },
    {
      "epoch": 95.91,
      "learning_rate": 7.5e-05,
      "loss": 0.9098,
      "step": 4700
    },
    {
      "epoch": 97.95,
      "learning_rate": 7.5e-05,
      "loss": 0.909,
      "step": 4800
    },
    {
      "epoch": 97.95,
      "eval_cer": 0.1212385012340139,
      "eval_loss": 0.4797753393650055,
      "eval_runtime": 116.9311,
      "eval_samples_per_second": 24.279,
      "eval_steps_per_second": 0.385,
      "eval_wer": 0.37645713280253373,
      "step": 4800
    },
    {
      "epoch": 99.99,
      "learning_rate": 7.5e-05,
      "loss": 0.8979,
      "step": 4900
    },
    {
      "epoch": 102.04,
      "learning_rate": 7.370797677261614e-05,
      "loss": 0.8993,
      "step": 5000
    },
    {
      "epoch": 104.08,
      "learning_rate": 7.22562652811736e-05,
      "loss": 0.8888,
      "step": 5100
    },
    {
      "epoch": 106.12,
      "learning_rate": 7.080455378973104e-05,
      "loss": 0.8788,
      "step": 5200
    },
    {
      "epoch": 106.12,
      "eval_cer": 0.11615885124523222,
      "eval_loss": 0.4906252324581146,
      "eval_runtime": 119.2237,
      "eval_samples_per_second": 23.812,
      "eval_steps_per_second": 0.377,
      "eval_wer": 0.3608410680508512,
      "step": 5200
    },
    {
      "epoch": 108.16,
      "learning_rate": 6.93528422982885e-05,
      "loss": 0.8701,
      "step": 5300
    },
    {
      "epoch": 110.2,
      "learning_rate": 6.790113080684596e-05,
      "loss": 0.8621,
      "step": 5400
    },
    {
      "epoch": 112.24,
      "learning_rate": 6.644941931540342e-05,
      "loss": 0.8525,
      "step": 5500
    },
    {
      "epoch": 114.28,
      "learning_rate": 6.499770782396088e-05,
      "loss": 0.8471,
      "step": 5600
    },
    {
      "epoch": 114.28,
      "eval_cer": 0.11655373569665695,
      "eval_loss": 0.4758923053741455,
      "eval_runtime": 117.4104,
      "eval_samples_per_second": 24.18,
      "eval_steps_per_second": 0.383,
      "eval_wer": 0.360357189988123,
      "step": 5600
    },
    {
      "epoch": 116.32,
      "learning_rate": 6.354599633251834e-05,
      "loss": 0.8442,
      "step": 5700
    },
    {
      "epoch": 118.36,
      "learning_rate": 6.209428484107578e-05,
      "loss": 0.8244,
      "step": 5800
    },
    {
      "epoch": 120.4,
      "learning_rate": 6.0642573349633246e-05,
      "loss": 0.8224,
      "step": 5900
    },
    {
      "epoch": 122.44,
      "learning_rate": 5.9190861858190705e-05,
      "loss": 0.8116,
      "step": 6000
    },
    {
      "epoch": 122.44,
      "eval_cer": 0.11757684541171191,
      "eval_loss": 0.5079970359802246,
      "eval_runtime": 118.6326,
      "eval_samples_per_second": 23.931,
      "eval_steps_per_second": 0.379,
      "eval_wer": 0.3627325913869705,
      "step": 6000
    },
    {
      "epoch": 124.48,
      "learning_rate": 5.773915036674816e-05,
      "loss": 0.807,
      "step": 6100
    },
    {
      "epoch": 126.53,
      "learning_rate": 5.6287438875305616e-05,
      "loss": 0.8011,
      "step": 6200
    },
    {
      "epoch": 128.57,
      "learning_rate": 5.4835727383863075e-05,
      "loss": 0.799,
      "step": 6300
    },
    {
      "epoch": 130.61,
      "learning_rate": 5.3384015892420534e-05,
      "loss": 0.7881,
      "step": 6400
    },
    {
      "epoch": 130.61,
      "eval_cer": 0.11347543190486875,
      "eval_loss": 0.4867645502090454,
      "eval_runtime": 118.7534,
      "eval_samples_per_second": 23.907,
      "eval_steps_per_second": 0.379,
      "eval_wer": 0.34892007214182025,
      "step": 6400
    },
    {
      "epoch": 132.65,
      "learning_rate": 5.193230440097799e-05,
      "loss": 0.7781,
      "step": 6500
    },
    {
      "epoch": 134.69,
      "learning_rate": 5.048059290953545e-05,
      "loss": 0.7685,
      "step": 6600
    },
    {
      "epoch": 136.73,
      "learning_rate": 4.902888141809291e-05,
      "loss": 0.7654,
      "step": 6700
    },
    {
      "epoch": 138.77,
      "learning_rate": 4.757716992665037e-05,
      "loss": 0.766,
      "step": 6800
    },
    {
      "epoch": 138.77,
      "eval_cer": 0.11357415301772493,
      "eval_loss": 0.49552106857299805,
      "eval_runtime": 117.4435,
      "eval_samples_per_second": 24.173,
      "eval_steps_per_second": 0.383,
      "eval_wer": 0.34922799454537456,
      "step": 6800
    },
    {
      "epoch": 140.81,
      "learning_rate": 4.6139975550122245e-05,
      "loss": 0.752,
      "step": 6900
    },
    {
      "epoch": 142.85,
      "learning_rate": 4.4688264058679704e-05,
      "loss": 0.7451,
      "step": 7000
    },
    {
      "epoch": 144.89,
      "learning_rate": 4.323655256723716e-05,
      "loss": 0.7398,
      "step": 7100
    },
    {
      "epoch": 146.93,
      "learning_rate": 4.178484107579462e-05,
      "loss": 0.7333,
      "step": 7200
    },
    {
      "epoch": 146.93,
      "eval_cer": 0.11247027148306035,
      "eval_loss": 0.5019353032112122,
      "eval_runtime": 118.907,
      "eval_samples_per_second": 23.876,
      "eval_steps_per_second": 0.378,
      "eval_wer": 0.3461487705098315,
      "step": 7200
    },
    {
      "epoch": 148.97,
      "learning_rate": 4.033312958435208e-05,
      "loss": 0.7255,
      "step": 7300
    },
    {
      "epoch": 151.02,
      "learning_rate": 3.888141809290954e-05,
      "loss": 0.7283,
      "step": 7400
    },
    {
      "epoch": 153.06,
      "learning_rate": 3.742970660146699e-05,
      "loss": 0.7149,
      "step": 7500
    },
    {
      "epoch": 155.1,
      "learning_rate": 3.597799511002445e-05,
      "loss": 0.709,
      "step": 7600
    },
    {
      "epoch": 155.1,
      "eval_cer": 0.11168050258021091,
      "eval_loss": 0.5083914995193481,
      "eval_runtime": 120.383,
      "eval_samples_per_second": 23.583,
      "eval_steps_per_second": 0.374,
      "eval_wer": 0.3468086042317336,
      "step": 7600
    },
    {
      "epoch": 157.14,
      "learning_rate": 3.452628361858191e-05,
      "loss": 0.7084,
      "step": 7700
    },
    {
      "epoch": 159.18,
      "learning_rate": 3.307457212713937e-05,
      "loss": 0.704,
      "step": 7800
    },
    {
      "epoch": 161.22,
      "learning_rate": 3.1622860635696827e-05,
      "loss": 0.698,
      "step": 7900
    },
    {
      "epoch": 163.26,
      "learning_rate": 3.0171149144254275e-05,
      "loss": 0.6911,
      "step": 8000
    },
    {
      "epoch": 163.26,
      "eval_cer": 0.11064841821853265,
      "eval_loss": 0.5144167542457581,
      "eval_runtime": 120.6954,
      "eval_samples_per_second": 23.522,
      "eval_steps_per_second": 0.373,
      "eval_wer": 0.34122201205296265,
      "step": 8000
    },
    {
      "epoch": 165.3,
      "learning_rate": 2.8719437652811734e-05,
      "loss": 0.6889,
      "step": 8100
    },
    {
      "epoch": 167.34,
      "learning_rate": 2.726772616136919e-05,
      "loss": 0.6863,
      "step": 8200
    },
    {
      "epoch": 169.38,
      "learning_rate": 2.581601466992665e-05,
      "loss": 0.6825,
      "step": 8300
    },
    {
      "epoch": 171.42,
      "learning_rate": 2.4364303178484107e-05,
      "loss": 0.6683,
      "step": 8400
    },
    {
      "epoch": 171.42,
      "eval_cer": 0.11165357864034103,
      "eval_loss": 0.5218909382820129,
      "eval_runtime": 119.5764,
      "eval_samples_per_second": 23.742,
      "eval_steps_per_second": 0.376,
      "eval_wer": 0.34091408964940834,
      "step": 8400
    },
    {
      "epoch": 173.46,
      "learning_rate": 2.2912591687041566e-05,
      "loss": 0.6627,
      "step": 8500
    },
    {
      "epoch": 175.51,
      "learning_rate": 2.1460880195599022e-05,
      "loss": 0.6655,
      "step": 8600
    },
    {
      "epoch": 177.55,
      "learning_rate": 2.000916870415648e-05,
      "loss": 0.6662,
      "step": 8700
    },
    {
      "epoch": 179.59,
      "learning_rate": 1.855745721271394e-05,
      "loss": 0.659,
      "step": 8800
    },
    {
      "epoch": 179.59,
      "eval_cer": 0.1095983845636078,
      "eval_loss": 0.5229596495628357,
      "eval_runtime": 119.167,
      "eval_samples_per_second": 23.824,
      "eval_steps_per_second": 0.378,
      "eval_wer": 0.3375709321251045,
      "step": 8800
    },
    {
      "epoch": 181.63,
      "learning_rate": 1.7105745721271395e-05,
      "loss": 0.6526,
      "step": 8900
    },
    {
      "epoch": 183.67,
      "learning_rate": 1.5654034229828854e-05,
      "loss": 0.654,
      "step": 9000
    },
    {
      "epoch": 185.71,
      "learning_rate": 1.4202322738386313e-05,
      "loss": 0.6481,
      "step": 9100
    },
    {
      "epoch": 187.75,
      "learning_rate": 1.275061124694377e-05,
      "loss": 0.6475,
      "step": 9200
    },
    {
      "epoch": 187.75,
      "eval_cer": 0.10967915638321742,
      "eval_loss": 0.5228886604309082,
      "eval_runtime": 118.4312,
      "eval_samples_per_second": 23.972,
      "eval_steps_per_second": 0.38,
      "eval_wer": 0.3397703778647781,
      "step": 9200
    },
    {
      "epoch": 189.79,
      "learning_rate": 1.1298899755501229e-05,
      "loss": 0.6476,
      "step": 9300
    },
    {
      "epoch": 191.83,
      "learning_rate": 9.847188264058686e-06,
      "loss": 0.6485,
      "step": 9400
    },
    {
      "epoch": 193.87,
      "learning_rate": 8.395476772616145e-06,
      "loss": 0.6449,
      "step": 9500
    },
    {
      "epoch": 195.91,
      "learning_rate": 6.943765281173602e-06,
      "loss": 0.6419,
      "step": 9600
    },
    {
      "epoch": 195.91,
      "eval_cer": 0.10843168050258022,
      "eval_loss": 0.520015299320221,
      "eval_runtime": 118.8519,
      "eval_samples_per_second": 23.887,
      "eval_steps_per_second": 0.379,
      "eval_wer": 0.33369990762327895,
      "step": 9600
    },
    {
      "epoch": 197.95,
      "learning_rate": 5.4920537897310516e-06,
      "loss": 0.627,
      "step": 9700
    },
    {
      "epoch": 199.99,
      "learning_rate": 4.054859413202941e-06,
      "loss": 0.6301,
      "step": 9800
    },
    {
      "epoch": 200.3,
      "step": 9815,
      "total_flos": 1.72528931544575e+20,
      "train_loss": 1.3263086397673363,
      "train_runtime": 56983.3895,
      "train_samples_per_second": 22.047,
      "train_steps_per_second": 0.172
    }
  ],
  "max_steps": 9815,
  "num_train_epochs": 201,
  "total_flos": 1.72528931544575e+20,
  "trial_name": null,
  "trial_params": null
}