{
  "best_metric": 0.09384384384384384,
  "best_model_checkpoint": "fine-w2v2base-bs16-ep100-lr2e-05-linguistic-rmsnorm-focal_ctc_a0.5_g0.5-0.05_10_0.004_40/checkpoint-2550",
  "epoch": 100.0,
  "eval_steps": 50,
  "global_step": 5300,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.94,
      "learning_rate": 7.547169811320755e-07,
      "loss": 1068.6505,
      "step": 50
    },
    {
      "epoch": 0.94,
      "eval_loss": 527.9987182617188,
      "eval_runtime": 3.3391,
      "eval_samples_per_second": 214.726,
      "eval_steps_per_second": 6.888,
      "eval_wer": 15.883967300633968,
      "step": 50
    },
    {
      "epoch": 1.89,
      "learning_rate": 2.6037735849056606e-06,
      "loss": 924.9194,
      "step": 100
    },
    {
      "epoch": 1.89,
      "eval_loss": 365.2997741699219,
      "eval_runtime": 2.3955,
      "eval_samples_per_second": 299.317,
      "eval_steps_per_second": 9.602,
      "eval_wer": 15.72897897897898,
      "step": 100
    },
    {
      "epoch": 2.83,
      "learning_rate": 4.49056603773585e-06,
      "loss": 268.1159,
      "step": 150
    },
    {
      "epoch": 2.83,
      "eval_loss": 45.30757141113281,
      "eval_runtime": 2.3434,
      "eval_samples_per_second": 305.972,
      "eval_steps_per_second": 9.815,
      "eval_wer": 1.0,
      "step": 150
    },
    {
      "epoch": 3.77,
      "learning_rate": 6.377358490566038e-06,
      "loss": 56.3914,
      "step": 200
    },
    {
      "epoch": 3.77,
      "eval_loss": 42.26533508300781,
      "eval_runtime": 2.4101,
      "eval_samples_per_second": 297.501,
      "eval_steps_per_second": 9.543,
      "eval_wer": 1.0,
      "step": 200
    },
    {
      "epoch": 4.72,
      "learning_rate": 8.264150943396228e-06,
      "loss": 54.5992,
      "step": 250
    },
    {
      "epoch": 4.72,
      "eval_loss": 41.22721481323242,
      "eval_runtime": 2.3201,
      "eval_samples_per_second": 309.032,
      "eval_steps_per_second": 9.913,
      "eval_wer": 1.0,
      "step": 250
    },
    {
      "epoch": 5.66,
      "learning_rate": 1.0150943396226416e-05,
      "loss": 52.7823,
      "step": 300
    },
    {
      "epoch": 5.66,
      "eval_loss": 40.21255111694336,
      "eval_runtime": 2.2946,
      "eval_samples_per_second": 312.473,
      "eval_steps_per_second": 10.024,
      "eval_wer": 1.0,
      "step": 300
    },
    {
      "epoch": 6.6,
      "learning_rate": 1.2037735849056605e-05,
      "loss": 51.1032,
      "step": 350
    },
    {
      "epoch": 6.6,
      "eval_loss": 39.62540817260742,
      "eval_runtime": 2.3293,
      "eval_samples_per_second": 307.813,
      "eval_steps_per_second": 9.874,
      "eval_wer": 1.0,
      "step": 350
    },
    {
      "epoch": 7.55,
      "learning_rate": 1.3924528301886793e-05,
      "loss": 49.2081,
      "step": 400
    },
    {
      "epoch": 7.55,
      "eval_loss": 38.79893493652344,
      "eval_runtime": 2.3075,
      "eval_samples_per_second": 310.729,
      "eval_steps_per_second": 9.968,
      "eval_wer": 1.0,
      "step": 400
    },
    {
      "epoch": 8.49,
      "learning_rate": 1.5811320754716985e-05,
      "loss": 48.3538,
      "step": 450
    },
    {
      "epoch": 8.49,
      "eval_loss": 38.57915115356445,
      "eval_runtime": 2.3577,
      "eval_samples_per_second": 304.109,
      "eval_steps_per_second": 9.755,
      "eval_wer": 1.0,
      "step": 450
    },
    {
      "epoch": 9.43,
      "learning_rate": 1.7698113207547173e-05,
      "loss": 48.8615,
      "step": 500
    },
    {
      "epoch": 9.43,
      "eval_loss": 38.462188720703125,
      "eval_runtime": 2.3275,
      "eval_samples_per_second": 308.057,
      "eval_steps_per_second": 9.882,
      "eval_wer": 1.0,
      "step": 500
    },
    {
      "epoch": 10.38,
      "learning_rate": 1.9584905660377362e-05,
      "loss": 48.1912,
      "step": 550
    },
    {
      "epoch": 10.38,
      "eval_loss": 38.14221954345703,
      "eval_runtime": 2.241,
      "eval_samples_per_second": 319.943,
      "eval_steps_per_second": 10.263,
      "eval_wer": 1.0,
      "step": 550
    },
    {
      "epoch": 11.32,
      "learning_rate": 1.9996701334124693e-05,
      "loss": 48.3589,
      "step": 600
    },
    {
      "epoch": 11.32,
      "eval_loss": 38.614410400390625,
      "eval_runtime": 2.3533,
      "eval_samples_per_second": 304.679,
      "eval_steps_per_second": 9.774,
      "eval_wer": 1.0,
      "step": 600
    },
    {
      "epoch": 12.26,
      "learning_rate": 1.9982825320106917e-05,
      "loss": 46.5985,
      "step": 650
    },
    {
      "epoch": 12.26,
      "eval_loss": 39.63941192626953,
      "eval_runtime": 2.2466,
      "eval_samples_per_second": 319.155,
      "eval_steps_per_second": 10.238,
      "eval_wer": 1.0472972972972974,
      "step": 650
    },
    {
      "epoch": 13.21,
      "learning_rate": 1.995812457240187e-05,
      "loss": 45.5769,
      "step": 700
    },
    {
      "epoch": 13.21,
      "eval_loss": 37.75796127319336,
      "eval_runtime": 2.3563,
      "eval_samples_per_second": 304.287,
      "eval_steps_per_second": 9.761,
      "eval_wer": 0.9992492492492493,
      "step": 700
    },
    {
      "epoch": 14.15,
      "learning_rate": 1.9922625874911624e-05,
      "loss": 44.1749,
      "step": 750
    },
    {
      "epoch": 14.15,
      "eval_loss": 36.06916427612305,
      "eval_runtime": 2.294,
      "eval_samples_per_second": 312.554,
      "eval_steps_per_second": 10.026,
      "eval_wer": 0.9990824157490824,
      "step": 750
    },
    {
      "epoch": 15.09,
      "learning_rate": 1.987636772014047e-05,
      "loss": 41.8932,
      "step": 800
    },
    {
      "epoch": 15.09,
      "eval_loss": 27.94037437438965,
      "eval_runtime": 2.4164,
      "eval_samples_per_second": 296.718,
      "eval_steps_per_second": 9.518,
      "eval_wer": 0.9315982649315983,
      "step": 800
    },
    {
      "epoch": 16.04,
      "learning_rate": 1.981940026745616e-05,
      "loss": 29.8551,
      "step": 850
    },
    {
      "epoch": 16.04,
      "eval_loss": 14.121084213256836,
      "eval_runtime": 2.3404,
      "eval_samples_per_second": 306.353,
      "eval_steps_per_second": 9.827,
      "eval_wer": 0.3929763096429763,
      "step": 850
    },
    {
      "epoch": 16.98,
      "learning_rate": 1.9751785288700255e-05,
      "loss": 16.9135,
      "step": 900
    },
    {
      "epoch": 16.98,
      "eval_loss": 7.982431411743164,
      "eval_runtime": 2.3673,
      "eval_samples_per_second": 302.882,
      "eval_steps_per_second": 9.716,
      "eval_wer": 0.22280613947280614,
      "step": 900
    },
    {
      "epoch": 17.92,
      "learning_rate": 1.9673596101206766e-05,
      "loss": 11.5569,
      "step": 950
    },
    {
      "epoch": 17.92,
      "eval_loss": 5.807270050048828,
      "eval_runtime": 2.351,
      "eval_samples_per_second": 304.982,
      "eval_steps_per_second": 9.783,
      "eval_wer": 0.16925258591925257,
      "step": 950
    },
    {
      "epoch": 18.87,
      "learning_rate": 1.9584917488301524e-05,
      "loss": 9.1965,
      "step": 1000
    },
    {
      "epoch": 18.87,
      "eval_loss": 4.689126491546631,
      "eval_runtime": 2.2997,
      "eval_samples_per_second": 311.783,
      "eval_steps_per_second": 10.001,
      "eval_wer": 0.15765765765765766,
      "step": 1000
    },
    {
      "epoch": 19.81,
      "learning_rate": 1.9485845607368606e-05,
      "loss": 7.6846,
      "step": 1050
    },
    {
      "epoch": 19.81,
      "eval_loss": 4.0937886238098145,
      "eval_runtime": 2.2421,
      "eval_samples_per_second": 319.793,
      "eval_steps_per_second": 10.258,
      "eval_wer": 0.1443943943943944,
      "step": 1050
    },
    {
      "epoch": 20.75,
      "learning_rate": 1.937648788558344e-05,
      "loss": 6.6186,
      "step": 1100
    },
    {
      "epoch": 20.75,
      "eval_loss": 3.7074310779571533,
      "eval_runtime": 2.2452,
      "eval_samples_per_second": 319.355,
      "eval_steps_per_second": 10.244,
      "eval_wer": 0.13371705038371706,
      "step": 1100
    },
    {
      "epoch": 21.7,
      "learning_rate": 1.925696290342571e-05,
      "loss": 6.1733,
      "step": 1150
    },
    {
      "epoch": 21.7,
      "eval_loss": 3.376906156539917,
      "eval_runtime": 2.3462,
      "eval_samples_per_second": 305.602,
      "eval_steps_per_second": 9.803,
      "eval_wer": 0.12787787787787788,
      "step": 1150
    },
    {
      "epoch": 22.64,
      "learning_rate": 1.912740026609828e-05,
      "loss": 5.5833,
      "step": 1200
    },
    {
      "epoch": 22.64,
      "eval_loss": 3.193333864212036,
      "eval_runtime": 2.4175,
      "eval_samples_per_second": 296.587,
      "eval_steps_per_second": 9.514,
      "eval_wer": 0.12879546212879547,
      "step": 1200
    },
    {
      "epoch": 23.58,
      "learning_rate": 1.8987940462991673e-05,
      "loss": 5.1097,
      "step": 1250
    },
    {
      "epoch": 23.58,
      "eval_loss": 3.078519582748413,
      "eval_runtime": 2.3455,
      "eval_samples_per_second": 305.697,
      "eval_steps_per_second": 9.806,
      "eval_wer": 0.12320653987320654,
      "step": 1250
    },
    {
      "epoch": 24.53,
      "learning_rate": 1.8838734715346398e-05,
      "loss": 4.8098,
      "step": 1300
    },
    {
      "epoch": 24.53,
      "eval_loss": 3.068678140640259,
      "eval_runtime": 2.2608,
      "eval_samples_per_second": 317.143,
      "eval_steps_per_second": 10.173,
      "eval_wer": 0.12103770437103771,
      "step": 1300
    },
    {
      "epoch": 25.47,
      "learning_rate": 1.867994481227837e-05,
      "loss": 4.784,
      "step": 1350
    },
    {
      "epoch": 25.47,
      "eval_loss": 2.777139186859131,
      "eval_runtime": 2.4013,
      "eval_samples_per_second": 298.593,
      "eval_steps_per_second": 9.578,
      "eval_wer": 0.11519853186519853,
      "step": 1350
    },
    {
      "epoch": 26.42,
      "learning_rate": 1.8511742935345198e-05,
      "loss": 4.3574,
      "step": 1400
    },
    {
      "epoch": 26.42,
      "eval_loss": 2.7346842288970947,
      "eval_runtime": 2.2397,
      "eval_samples_per_second": 320.13,
      "eval_steps_per_second": 10.269,
      "eval_wer": 0.1200367033700367,
      "step": 1400
    },
    {
      "epoch": 27.36,
      "learning_rate": 1.8334311471843574e-05,
      "loss": 4.2972,
      "step": 1450
    },
    {
      "epoch": 27.36,
      "eval_loss": 2.685316324234009,
      "eval_runtime": 2.3028,
      "eval_samples_per_second": 311.361,
      "eval_steps_per_second": 9.988,
      "eval_wer": 0.11469803136469803,
      "step": 1450
    },
    {
      "epoch": 28.3,
      "learning_rate": 1.814784281704023e-05,
      "loss": 4.1072,
      "step": 1500
    },
    {
      "epoch": 28.3,
      "eval_loss": 2.5679965019226074,
      "eval_runtime": 2.294,
      "eval_samples_per_second": 312.557,
      "eval_steps_per_second": 10.026,
      "eval_wer": 0.11845178511845178,
      "step": 1500
    },
    {
      "epoch": 29.25,
      "learning_rate": 1.7952539165550863e-05,
      "loss": 3.9651,
      "step": 1550
    },
    {
      "epoch": 29.25,
      "eval_loss": 2.5938329696655273,
      "eval_runtime": 2.4099,
      "eval_samples_per_second": 297.522,
      "eval_steps_per_second": 9.544,
      "eval_wer": 0.1200367033700367,
      "step": 1550
    },
    {
      "epoch": 30.19,
      "learning_rate": 1.7748612292093336e-05,
      "loss": 4.0325,
      "step": 1600
    },
    {
      "epoch": 30.19,
      "eval_loss": 2.5324084758758545,
      "eval_runtime": 2.2799,
      "eval_samples_per_second": 314.482,
      "eval_steps_per_second": 10.088,
      "eval_wer": 0.1180347013680347,
      "step": 1600
    },
    {
      "epoch": 31.13,
      "learning_rate": 1.753628332185275e-05,
      "loss": 3.6586,
      "step": 1650
    },
    {
      "epoch": 31.13,
      "eval_loss": 2.584696054458618,
      "eval_runtime": 2.4425,
      "eval_samples_per_second": 293.556,
      "eval_steps_per_second": 9.417,
      "eval_wer": 0.11127794461127795,
      "step": 1650
    },
    {
      "epoch": 32.08,
      "learning_rate": 1.731578249070756e-05,
      "loss": 3.7213,
      "step": 1700
    },
    {
      "epoch": 32.08,
      "eval_loss": 2.5886008739471436,
      "eval_runtime": 2.2464,
      "eval_samples_per_second": 319.179,
      "eval_steps_per_second": 10.239,
      "eval_wer": 0.11161161161161161,
      "step": 1700
    },
    {
      "epoch": 33.02,
      "learning_rate": 1.7087348895576564e-05,
      "loss": 3.4746,
      "step": 1750
    },
    {
      "epoch": 33.02,
      "eval_loss": 2.428494453430176,
      "eval_runtime": 2.4929,
      "eval_samples_per_second": 287.612,
      "eval_steps_per_second": 9.226,
      "eval_wer": 0.10051718385051718,
      "step": 1750
    },
    {
      "epoch": 33.96,
      "learning_rate": 1.68512302351576e-05,
      "loss": 3.3572,
      "step": 1800
    },
    {
      "epoch": 33.96,
      "eval_loss": 2.4606618881225586,
      "eval_runtime": 2.2482,
      "eval_samples_per_second": 318.92,
      "eval_steps_per_second": 10.23,
      "eval_wer": 0.10727394060727394,
      "step": 1800
    },
    {
      "epoch": 34.91,
      "learning_rate": 1.6607682541338998e-05,
      "loss": 3.2202,
      "step": 1850
    },
    {
      "epoch": 34.91,
      "eval_loss": 2.4459075927734375,
      "eval_runtime": 2.406,
      "eval_samples_per_second": 298.009,
      "eval_steps_per_second": 9.56,
      "eval_wer": 0.11027694361027694,
      "step": 1850
    },
    {
      "epoch": 35.85,
      "learning_rate": 1.6356969901575094e-05,
      "loss": 3.2437,
      "step": 1900
    },
    {
      "epoch": 35.85,
      "eval_loss": 2.3630285263061523,
      "eval_runtime": 2.2948,
      "eval_samples_per_second": 312.446,
      "eval_steps_per_second": 10.023,
      "eval_wer": 0.10268601935268602,
      "step": 1900
    },
    {
      "epoch": 36.79,
      "learning_rate": 1.6099364172526732e-05,
      "loss": 3.1303,
      "step": 1950
    },
    {
      "epoch": 36.79,
      "eval_loss": 2.32814884185791,
      "eval_runtime": 2.3311,
      "eval_samples_per_second": 307.582,
      "eval_steps_per_second": 9.867,
      "eval_wer": 0.10251918585251918,
      "step": 1950
    },
    {
      "epoch": 37.74,
      "learning_rate": 1.583514468527744e-05,
      "loss": 3.0037,
      "step": 2000
    },
    {
      "epoch": 37.74,
      "eval_loss": 2.3129348754882812,
      "eval_runtime": 2.4145,
      "eval_samples_per_second": 296.962,
      "eval_steps_per_second": 9.526,
      "eval_wer": 0.10185185185185185,
      "step": 2000
    },
    {
      "epoch": 38.68,
      "learning_rate": 1.5564597942444743e-05,
      "loss": 3.0523,
      "step": 2050
    },
    {
      "epoch": 38.68,
      "eval_loss": 2.2962000370025635,
      "eval_runtime": 2.4793,
      "eval_samples_per_second": 289.191,
      "eval_steps_per_second": 9.277,
      "eval_wer": 0.09884884884884886,
      "step": 2050
    },
    {
      "epoch": 39.62,
      "learning_rate": 1.5288017307515142e-05,
      "loss": 2.8943,
      "step": 2100
    },
    {
      "epoch": 39.62,
      "eval_loss": 2.323789119720459,
      "eval_runtime": 2.2485,
      "eval_samples_per_second": 318.884,
      "eval_steps_per_second": 10.229,
      "eval_wer": 0.1021021021021021,
      "step": 2100
    },
    {
      "epoch": 40.57,
      "learning_rate": 1.500570268673965e-05,
      "loss": 2.8502,
      "step": 2150
    },
    {
      "epoch": 40.57,
      "eval_loss": 2.3549041748046875,
      "eval_runtime": 2.4526,
      "eval_samples_per_second": 292.339,
      "eval_steps_per_second": 9.378,
      "eval_wer": 0.10435435435435435,
      "step": 2150
    },
    {
      "epoch": 41.51,
      "learning_rate": 1.4717960203934704e-05,
      "loss": 2.7045,
      "step": 2200
    },
    {
      "epoch": 41.51,
      "eval_loss": 2.367952346801758,
      "eval_runtime": 2.2393,
      "eval_samples_per_second": 320.195,
      "eval_steps_per_second": 10.271,
      "eval_wer": 0.10176843510176843,
      "step": 2200
    },
    {
      "epoch": 42.45,
      "learning_rate": 1.4425101868541228e-05,
      "loss": 2.7291,
      "step": 2250
    },
    {
      "epoch": 42.45,
      "eval_loss": 2.4171645641326904,
      "eval_runtime": 2.2967,
      "eval_samples_per_second": 312.191,
      "eval_steps_per_second": 10.015,
      "eval_wer": 0.11286286286286286,
      "step": 2250
    },
    {
      "epoch": 43.4,
      "learning_rate": 1.412744523730163e-05,
      "loss": 2.6162,
      "step": 2300
    },
    {
      "epoch": 43.4,
      "eval_loss": 2.321629524230957,
      "eval_runtime": 2.3019,
      "eval_samples_per_second": 311.476,
      "eval_steps_per_second": 9.992,
      "eval_wer": 0.10176843510176843,
      "step": 2300
    },
    {
      "epoch": 44.34,
      "learning_rate": 1.3825313069921713e-05,
      "loss": 2.5643,
      "step": 2350
    },
    {
      "epoch": 44.34,
      "eval_loss": 2.2663228511810303,
      "eval_runtime": 2.4493,
      "eval_samples_per_second": 292.741,
      "eval_steps_per_second": 9.391,
      "eval_wer": 0.09793126459793126,
      "step": 2350
    },
    {
      "epoch": 45.28,
      "learning_rate": 1.3519032979090816e-05,
      "loss": 2.5842,
      "step": 2400
    },
    {
      "epoch": 45.28,
      "eval_loss": 2.240830898284912,
      "eval_runtime": 2.2616,
      "eval_samples_per_second": 317.033,
      "eval_steps_per_second": 10.17,
      "eval_wer": 0.0985985985985986,
      "step": 2400
    },
    {
      "epoch": 46.23,
      "learning_rate": 1.3208937075239663e-05,
      "loss": 2.4498,
      "step": 2450
    },
    {
      "epoch": 46.23,
      "eval_loss": 2.2695207595825195,
      "eval_runtime": 2.3803,
      "eval_samples_per_second": 301.222,
      "eval_steps_per_second": 9.663,
      "eval_wer": 0.10168501835168502,
      "step": 2450
    },
    {
      "epoch": 47.17,
      "learning_rate": 1.289536160642119e-05,
      "loss": 2.4177,
      "step": 2500
    },
    {
      "epoch": 47.17,
      "eval_loss": 2.202859878540039,
      "eval_runtime": 2.2399,
      "eval_samples_per_second": 320.105,
      "eval_steps_per_second": 10.268,
      "eval_wer": 0.09801468134801468,
      "step": 2500
    },
    {
      "epoch": 48.11,
      "learning_rate": 1.2578646593704786e-05,
      "loss": 2.3297,
      "step": 2550
    },
    {
      "epoch": 48.11,
      "eval_loss": 2.225446939468384,
      "eval_runtime": 2.3113,
      "eval_samples_per_second": 310.217,
      "eval_steps_per_second": 9.951,
      "eval_wer": 0.09384384384384384,
      "step": 2550
    },
    {
      "epoch": 49.06,
      "learning_rate": 1.2259135462479306e-05,
      "loss": 2.3637,
      "step": 2600
    },
    {
      "epoch": 49.06,
      "eval_loss": 2.2550511360168457,
      "eval_runtime": 2.3163,
      "eval_samples_per_second": 309.542,
      "eval_steps_per_second": 9.93,
      "eval_wer": 0.1011011011011011,
      "step": 2600
    },
    {
      "epoch": 50.0,
      "learning_rate": 1.1937174670064665e-05,
      "loss": 2.2528,
      "step": 2650
    },
    {
      "epoch": 50.0,
      "eval_loss": 2.235013723373413,
      "eval_runtime": 2.297,
      "eval_samples_per_second": 312.152,
      "eval_steps_per_second": 10.013,
      "eval_wer": 0.10118451785118451,
      "step": 2650
    },
    {
      "epoch": 50.94,
      "learning_rate": 1.1613113330035816e-05,
      "loss": 2.2221,
      "step": 2700
    },
    {
      "epoch": 50.94,
      "eval_loss": 2.2253198623657227,
      "eval_runtime": 2.4122,
      "eval_samples_per_second": 297.24,
      "eval_steps_per_second": 9.535,
      "eval_wer": 0.09676343009676343,
      "step": 2700
    },
    {
      "epoch": 51.89,
      "learning_rate": 1.1287302833666442e-05,
      "loss": 2.3083,
      "step": 2750
    },
    {
      "epoch": 51.89,
      "eval_loss": 2.242628335952759,
      "eval_runtime": 2.4199,
      "eval_samples_per_second": 296.292,
      "eval_steps_per_second": 9.504,
      "eval_wer": 0.09584584584584585,
      "step": 2750
    },
    {
      "epoch": 52.83,
      "learning_rate": 1.0960096468902856e-05,
      "loss": 2.0585,
      "step": 2800
    },
    {
      "epoch": 52.83,
      "eval_loss": 2.216921806335449,
      "eval_runtime": 2.2934,
      "eval_samples_per_second": 312.636,
      "eval_steps_per_second": 10.029,
      "eval_wer": 0.09718051384718052,
      "step": 2800
    },
    {
      "epoch": 53.77,
      "learning_rate": 1.0631849037281267e-05,
      "loss": 2.2349,
      "step": 2850
    },
    {
      "epoch": 53.77,
      "eval_loss": 2.215132236480713,
      "eval_runtime": 2.3656,
      "eval_samples_per_second": 303.097,
      "eval_steps_per_second": 9.723,
      "eval_wer": 0.10035035035035035,
      "step": 2850
    },
    {
      "epoch": 54.72,
      "learning_rate": 1.0302916469203831e-05,
      "loss": 2.1969,
      "step": 2900
    },
    {
      "epoch": 54.72,
      "eval_loss": 2.2561628818511963,
      "eval_runtime": 2.2347,
      "eval_samples_per_second": 320.851,
      "eval_steps_per_second": 10.292,
      "eval_wer": 0.10235235235235235,
      "step": 2900
    },
    {
      "epoch": 55.66,
      "learning_rate": 9.973655437990618e-06,
      "loss": 2.0415,
      "step": 2950
    },
    {
      "epoch": 55.66,
      "eval_loss": 2.2862448692321777,
      "eval_runtime": 2.3242,
      "eval_samples_per_second": 308.491,
      "eval_steps_per_second": 9.896,
      "eval_wer": 0.10268601935268602,
      "step": 2950
    },
    {
      "epoch": 56.6,
      "learning_rate": 9.644422973125977e-06,
      "loss": 2.0126,
      "step": 3000
    },
    {
      "epoch": 56.6,
      "eval_loss": 2.2166805267333984,
      "eval_runtime": 2.2438,
      "eval_samples_per_second": 319.545,
      "eval_steps_per_second": 10.25,
      "eval_wer": 0.10151818485151819,
      "step": 3000
    },
    {
      "epoch": 57.55,
      "learning_rate": 9.31557607311876e-06,
      "loss": 2.1,
      "step": 3050
    },
    {
      "epoch": 57.55,
      "eval_loss": 2.2359628677368164,
      "eval_runtime": 2.3078,
      "eval_samples_per_second": 310.684,
      "eval_steps_per_second": 9.966,
      "eval_wer": 0.10243576910243576,
      "step": 3050
    },
    {
      "epoch": 58.49,
      "learning_rate": 8.987471318396079e-06,
      "loss": 2.0739,
      "step": 3100
    },
    {
      "epoch": 58.49,
      "eval_loss": 2.219820976257324,
      "eval_runtime": 2.2533,
      "eval_samples_per_second": 318.193,
      "eval_steps_per_second": 10.207,
      "eval_wer": 0.10560560560560561,
      "step": 3100
    },
    {
      "epoch": 59.43,
      "learning_rate": 8.660464484650442e-06,
      "loss": 1.9875,
      "step": 3150
    },
    {
      "epoch": 59.43,
      "eval_loss": 2.1715681552886963,
      "eval_runtime": 2.3021,
      "eval_samples_per_second": 311.449,
      "eval_steps_per_second": 9.991,
      "eval_wer": 0.09868201534868201,
      "step": 3150
    },
    {
      "epoch": 60.38,
      "learning_rate": 8.334910157059459e-06,
      "loss": 2.0259,
      "step": 3200
    },
    {
      "epoch": 60.38,
      "eval_loss": 2.214280128479004,
      "eval_runtime": 2.2789,
      "eval_samples_per_second": 314.622,
      "eval_steps_per_second": 10.092,
      "eval_wer": 0.09993326659993326,
      "step": 3200
    },
    {
      "epoch": 61.32,
      "learning_rate": 8.011161345796495e-06,
      "loss": 1.8519,
      "step": 3250
    },
    {
      "epoch": 61.32,
      "eval_loss": 2.1836636066436768,
      "eval_runtime": 2.345,
      "eval_samples_per_second": 305.758,
      "eval_steps_per_second": 9.808,
      "eval_wer": 0.09584584584584585,
      "step": 3250
    },
    {
      "epoch": 62.26,
      "learning_rate": 7.689569103249139e-06,
      "loss": 1.9733,
      "step": 3300
    },
    {
      "epoch": 62.26,
      "eval_loss": 2.1865181922912598,
      "eval_runtime": 2.3454,
      "eval_samples_per_second": 305.699,
      "eval_steps_per_second": 9.806,
      "eval_wer": 0.10076743410076744,
      "step": 3300
    },
    {
      "epoch": 63.21,
      "learning_rate": 7.3704821433605685e-06,
      "loss": 1.8496,
      "step": 3350
    },
    {
      "epoch": 63.21,
      "eval_loss": 2.2044992446899414,
      "eval_runtime": 2.4785,
      "eval_samples_per_second": 289.29,
      "eval_steps_per_second": 9.28,
      "eval_wer": 0.10535535535535535,
      "step": 3350
    },
    {
      "epoch": 64.15,
      "learning_rate": 7.054246463506596e-06,
      "loss": 1.9354,
      "step": 3400
    },
    {
      "epoch": 64.15,
      "eval_loss": 2.1783363819122314,
      "eval_runtime": 2.2431,
      "eval_samples_per_second": 319.643,
      "eval_steps_per_second": 10.254,
      "eval_wer": 0.10018351685018352,
      "step": 3400
    },
    {
      "epoch": 65.09,
      "learning_rate": 6.741204969318343e-06,
      "loss": 1.8247,
      "step": 3450
    },
    {
      "epoch": 65.09,
      "eval_loss": 2.166977882385254,
      "eval_runtime": 2.2716,
      "eval_samples_per_second": 315.638,
      "eval_steps_per_second": 10.125,
      "eval_wer": 0.09893226559893227,
      "step": 3450
    },
    {
      "epoch": 66.04,
      "learning_rate": 6.43169710285745e-06,
      "loss": 1.8418,
      "step": 3500
    },
    {
      "epoch": 66.04,
      "eval_loss": 2.182258129119873,
      "eval_runtime": 2.4022,
      "eval_samples_per_second": 298.482,
      "eval_steps_per_second": 9.575,
      "eval_wer": 0.09926593259926593,
      "step": 3500
    },
    {
      "epoch": 66.98,
      "learning_rate": 6.126058474546936e-06,
      "loss": 1.8259,
      "step": 3550
    },
    {
      "epoch": 66.98,
      "eval_loss": 2.1875205039978027,
      "eval_runtime": 2.3977,
      "eval_samples_per_second": 299.039,
      "eval_steps_per_second": 9.593,
      "eval_wer": 0.09901568234901569,
      "step": 3550
    },
    {
      "epoch": 67.92,
      "learning_rate": 5.82462049925683e-06,
      "loss": 1.8458,
      "step": 3600
    },
    {
      "epoch": 67.92,
      "eval_loss": 2.2048122882843018,
      "eval_runtime": 2.2439,
      "eval_samples_per_second": 319.528,
      "eval_steps_per_second": 10.25,
      "eval_wer": 0.10001668335001668,
      "step": 3600
    },
    {
      "epoch": 68.87,
      "learning_rate": 5.527710036939207e-06,
      "loss": 1.7796,
      "step": 3650
    },
    {
      "epoch": 68.87,
      "eval_loss": 2.201897382736206,
      "eval_runtime": 2.3586,
      "eval_samples_per_second": 303.996,
      "eval_steps_per_second": 9.752,
      "eval_wer": 0.09751418084751418,
      "step": 3650
    },
    {
      "epoch": 69.81,
      "learning_rate": 5.235649038202294e-06,
      "loss": 1.7931,
      "step": 3700
    },
    {
      "epoch": 69.81,
      "eval_loss": 2.1672749519348145,
      "eval_runtime": 2.3721,
      "eval_samples_per_second": 302.258,
      "eval_steps_per_second": 9.696,
      "eval_wer": 0.09551217884551218,
      "step": 3700
    },
    {
      "epoch": 70.75,
      "learning_rate": 4.948754195207908e-06,
      "loss": 1.789,
      "step": 3750
    },
    {
      "epoch": 70.75,
      "eval_loss": 2.1924376487731934,
      "eval_runtime": 2.242,
      "eval_samples_per_second": 319.808,
      "eval_steps_per_second": 10.259,
      "eval_wer": 0.09851518184851518,
      "step": 3750
    },
    {
      "epoch": 71.7,
      "learning_rate": 4.6673365982708805e-06,
      "loss": 1.8166,
      "step": 3800
    },
    {
      "epoch": 71.7,
      "eval_loss": 2.183880567550659,
      "eval_runtime": 2.3389,
      "eval_samples_per_second": 306.558,
      "eval_steps_per_second": 9.834,
      "eval_wer": 0.09642976309642977,
      "step": 3800
    },
    {
      "epoch": 72.64,
      "learning_rate": 4.3917013985327075e-06,
      "loss": 1.692,
      "step": 3850
    },
    {
      "epoch": 72.64,
      "eval_loss": 2.1770949363708496,
      "eval_runtime": 2.2882,
      "eval_samples_per_second": 313.348,
      "eval_steps_per_second": 10.052,
      "eval_wer": 0.09501167834501167,
      "step": 3850
    },
    {
      "epoch": 73.58,
      "learning_rate": 4.12214747707527e-06,
      "loss": 1.6898,
      "step": 3900
    },
    {
      "epoch": 73.58,
      "eval_loss": 2.162118911743164,
      "eval_runtime": 2.4655,
      "eval_samples_per_second": 290.811,
      "eval_steps_per_second": 9.329,
      "eval_wer": 0.09442776109442776,
      "step": 3900
    },
    {
      "epoch": 74.53,
      "learning_rate": 3.8589671208334414e-06,
      "loss": 1.5916,
      "step": 3950
    },
    {
      "epoch": 74.53,
      "eval_loss": 2.171839714050293,
      "eval_runtime": 2.2934,
      "eval_samples_per_second": 312.641,
      "eval_steps_per_second": 10.029,
      "eval_wer": 0.09734734734734735,
      "step": 3950
    },
    {
      "epoch": 75.47,
      "learning_rate": 3.6024457056579186e-06,
      "loss": 1.7778,
      "step": 4000
    },
    {
      "epoch": 75.47,
      "eval_loss": 2.1617255210876465,
      "eval_runtime": 2.3683,
      "eval_samples_per_second": 302.752,
      "eval_steps_per_second": 9.712,
      "eval_wer": 0.09734734734734735,
      "step": 4000
    },
    {
      "epoch": 76.42,
      "learning_rate": 3.352861386871993e-06,
      "loss": 1.6884,
      "step": 4050
    },
    {
      "epoch": 76.42,
      "eval_loss": 2.1565628051757812,
      "eval_runtime": 2.2914,
      "eval_samples_per_second": 312.914,
      "eval_steps_per_second": 10.038,
      "eval_wer": 0.09818151484818151,
      "step": 4050
    },
    {
      "epoch": 77.36,
      "learning_rate": 3.1104847976578332e-06,
      "loss": 1.7182,
      "step": 4100
    },
    {
      "epoch": 77.36,
      "eval_loss": 2.1698944568634033,
      "eval_runtime": 2.4574,
      "eval_samples_per_second": 291.777,
      "eval_steps_per_second": 9.36,
      "eval_wer": 0.09676343009676343,
      "step": 4100
    },
    {
      "epoch": 78.3,
      "learning_rate": 2.8755787555992578e-06,
      "loss": 1.6774,
      "step": 4150
    },
    {
      "epoch": 78.3,
      "eval_loss": 2.184875011444092,
      "eval_runtime": 2.4088,
      "eval_samples_per_second": 297.655,
      "eval_steps_per_second": 9.548,
      "eval_wer": 0.09642976309642977,
      "step": 4150
    },
    {
      "epoch": 79.25,
      "learning_rate": 2.6483979776992406e-06,
      "loss": 1.5921,
      "step": 4200
    },
    {
      "epoch": 79.25,
      "eval_loss": 2.1785097122192383,
      "eval_runtime": 2.2929,
      "eval_samples_per_second": 312.709,
      "eval_steps_per_second": 10.031,
      "eval_wer": 0.09617951284617951,
      "step": 4200
    },
    {
      "epoch": 80.19,
      "learning_rate": 2.429188804181195e-06,
      "loss": 1.7108,
      "step": 4250
    },
    {
      "epoch": 80.19,
      "eval_loss": 2.1641573905944824,
      "eval_runtime": 2.4247,
      "eval_samples_per_second": 295.704,
      "eval_steps_per_second": 9.486,
      "eval_wer": 0.0980980980980981,
      "step": 4250
    },
    {
      "epoch": 81.13,
      "learning_rate": 2.2181889313734763e-06,
      "loss": 1.7039,
      "step": 4300
    },
    {
      "epoch": 81.13,
      "eval_loss": 2.18363094329834,
      "eval_runtime": 2.3288,
      "eval_samples_per_second": 307.881,
      "eval_steps_per_second": 9.876,
      "eval_wer": 0.09968301634968302,
      "step": 4300
    },
    {
      "epoch": 82.08,
      "learning_rate": 2.0156271539667517e-06,
      "loss": 1.6068,
      "step": 4350
    },
    {
      "epoch": 82.08,
      "eval_loss": 2.1923670768737793,
      "eval_runtime": 2.2451,
      "eval_samples_per_second": 319.365,
      "eval_steps_per_second": 10.245,
      "eval_wer": 0.10018351685018352,
      "step": 4350
    },
    {
      "epoch": 83.02,
      "learning_rate": 1.8217231169237837e-06,
      "loss": 1.6267,
      "step": 4400
    },
    {
      "epoch": 83.02,
      "eval_loss": 2.1807827949523926,
      "eval_runtime": 2.2451,
      "eval_samples_per_second": 319.363,
      "eval_steps_per_second": 10.245,
      "eval_wer": 0.09793126459793126,
      "step": 4400
    },
    {
      "epoch": 83.96,
      "learning_rate": 1.6366870773105415e-06,
      "loss": 1.6209,
      "step": 4450
    },
    {
      "epoch": 83.96,
      "eval_loss": 2.1807706356048584,
      "eval_runtime": 2.2844,
      "eval_samples_per_second": 313.863,
      "eval_steps_per_second": 10.068,
      "eval_wer": 0.09759759759759759,
      "step": 4450
    },
    {
      "epoch": 84.91,
      "learning_rate": 1.460719676306962e-06,
      "loss": 1.6989,
      "step": 4500
    },
    {
      "epoch": 84.91,
      "eval_loss": 2.1661221981048584,
      "eval_runtime": 2.4206,
      "eval_samples_per_second": 296.21,
      "eval_steps_per_second": 9.502,
      "eval_wer": 0.09759759759759759,
      "step": 4500
    },
    {
      "epoch": 85.85,
      "learning_rate": 1.294011721644568e-06,
      "loss": 1.6126,
      "step": 4550
    },
    {
      "epoch": 85.85,
      "eval_loss": 2.1738200187683105,
      "eval_runtime": 2.2807,
      "eval_samples_per_second": 314.376,
      "eval_steps_per_second": 10.085,
      "eval_wer": 0.09884884884884886,
      "step": 4550
    },
    {
      "epoch": 86.79,
      "learning_rate": 1.1367439807068337e-06,
      "loss": 1.6623,
      "step": 4600
    },
    {
      "epoch": 86.79,
      "eval_loss": 2.1695382595062256,
      "eval_runtime": 2.2977,
      "eval_samples_per_second": 312.057,
      "eval_steps_per_second": 10.01,
      "eval_wer": 0.09793126459793126,
      "step": 4600
    },
    {
      "epoch": 87.74,
      "learning_rate": 9.890869845166518e-07,
      "loss": 1.637,
      "step": 4650
    },
    {
      "epoch": 87.74,
      "eval_loss": 2.1701598167419434,
      "eval_runtime": 2.291,
      "eval_samples_per_second": 312.966,
      "eval_steps_per_second": 10.039,
      "eval_wer": 0.09893226559893227,
      "step": 4650
    },
    {
      "epoch": 88.68,
      "learning_rate": 8.512008428234775e-07,
      "loss": 1.63,
      "step": 4700
    },
    {
      "epoch": 88.68,
      "eval_loss": 2.163140296936035,
      "eval_runtime": 2.3179,
      "eval_samples_per_second": 309.332,
      "eval_steps_per_second": 9.923,
      "eval_wer": 0.09734734734734735,
      "step": 4700
    },
    {
      "epoch": 89.62,
      "learning_rate": 7.232350704906032e-07,
      "loss": 1.6153,
      "step": 4750
    },
    {
      "epoch": 89.62,
      "eval_loss": 2.165863037109375,
      "eval_runtime": 2.2389,
      "eval_samples_per_second": 320.252,
      "eval_steps_per_second": 10.273,
      "eval_wer": 0.09851518184851518,
      "step": 4750
    },
    {
      "epoch": 90.57,
      "learning_rate": 6.053284253708547e-07,
      "loss": 1.4989,
      "step": 4800
    },
    {
      "epoch": 90.57,
      "eval_loss": 2.169069528579712,
      "eval_runtime": 2.3639,
      "eval_samples_per_second": 303.318,
      "eval_steps_per_second": 9.73,
      "eval_wer": 0.0990990990990991,
      "step": 4800
    },
    {
      "epoch": 91.51,
      "learning_rate": 4.976087578465116e-07,
      "loss": 1.7316,
      "step": 4850
    },
    {
      "epoch": 91.51,
      "eval_loss": 2.1687872409820557,
      "eval_runtime": 2.4114,
      "eval_samples_per_second": 297.337,
      "eval_steps_per_second": 9.538,
      "eval_wer": 0.0985985985985986,
      "step": 4850
    },
    {
      "epoch": 92.45,
      "learning_rate": 4.0019287219656646e-07,
      "loss": 1.4623,
      "step": 4900
    },
    {
      "epoch": 92.45,
      "eval_loss": 2.163541316986084,
      "eval_runtime": 2.4367,
      "eval_samples_per_second": 294.251,
      "eval_steps_per_second": 9.439,
      "eval_wer": 0.09801468134801468,
      "step": 4900
    },
    {
      "epoch": 93.4,
      "learning_rate": 3.1318639994168176e-07,
      "loss": 1.6932,
      "step": 4950
    },
    {
      "epoch": 93.4,
      "eval_loss": 2.1671087741851807,
      "eval_runtime": 2.3756,
      "eval_samples_per_second": 301.824,
      "eval_steps_per_second": 9.682,
      "eval_wer": 0.0985985985985986,
      "step": 4950
    },
    {
      "epoch": 94.34,
      "learning_rate": 2.366836853041621e-07,
      "loss": 1.5762,
      "step": 5000
    },
    {
      "epoch": 94.34,
      "eval_loss": 2.167776107788086,
      "eval_runtime": 2.3981,
      "eval_samples_per_second": 298.987,
      "eval_steps_per_second": 9.591,
      "eval_wer": 0.09901568234901569,
      "step": 5000
    },
    {
      "epoch": 95.28,
      "learning_rate": 1.7076768290714806e-07,
      "loss": 1.5346,
      "step": 5050
    },
    {
      "epoch": 95.28,
      "eval_loss": 2.1653971672058105,
      "eval_runtime": 2.4313,
      "eval_samples_per_second": 294.9,
      "eval_steps_per_second": 9.46,
      "eval_wer": 0.09843176509843177,
      "step": 5050
    },
    {
      "epoch": 96.23,
      "learning_rate": 1.1550986782395857e-07,
      "loss": 1.6015,
      "step": 5100
    },
    {
      "epoch": 96.23,
      "eval_loss": 2.166715383529663,
      "eval_runtime": 2.2602,
      "eval_samples_per_second": 317.23,
      "eval_steps_per_second": 10.176,
      "eval_wer": 0.0985985985985986,
      "step": 5100
    },
    {
      "epoch": 97.17,
      "learning_rate": 7.097015807511542e-08,
      "loss": 1.5609,
      "step": 5150
    },
    {
      "epoch": 97.17,
      "eval_loss": 2.1653175354003906,
      "eval_runtime": 2.2916,
      "eval_samples_per_second": 312.882,
      "eval_steps_per_second": 10.037,
      "eval_wer": 0.09818151484818151,
      "step": 5150
    },
    {
      "epoch": 98.11,
      "learning_rate": 3.71968496571018e-08,
      "loss": 1.6414,
      "step": 5200
    },
    {
      "epoch": 98.11,
      "eval_loss": 2.1648495197296143,
      "eval_runtime": 2.2488,
      "eval_samples_per_second": 318.838,
      "eval_steps_per_second": 10.228,
      "eval_wer": 0.09826493159826494,
      "step": 5200
    },
    {
      "epoch": 99.06,
      "learning_rate": 1.4579919805198795e-08,
      "loss": 1.581,
      "step": 5250
    },
    {
      "epoch": 99.06,
      "eval_loss": 2.166619300842285,
      "eval_runtime": 2.3312,
      "eval_samples_per_second": 307.572,
      "eval_steps_per_second": 9.866,
      "eval_wer": 0.09868201534868201,
      "step": 5250
    },
    {
      "epoch": 100.0,
      "learning_rate": 2.2208379630039858e-09,
      "loss": 1.6469,
      "step": 5300
    },
    {
      "epoch": 100.0,
      "eval_loss": 2.1651339530944824,
      "eval_runtime": 2.2476,
      "eval_samples_per_second": 319.01,
      "eval_steps_per_second": 10.233,
      "eval_wer": 0.09851518184851518,
      "step": 5300
    },
    {
      "epoch": 100.0,
      "step": 5300,
      "total_flos": 8.039557232590848e+16,
      "train_loss": 30.018039647228314,
      "train_runtime": 4054.7018,
      "train_samples_per_second": 83.459,
      "train_steps_per_second": 1.307
    }
  ],
  "logging_steps": 50,
  "max_steps": 5300,
  "num_train_epochs": 100,
  "save_steps": 50,
  "total_flos": 8.039557232590848e+16,
  "trial_name": null,
  "trial_params": null
}