{
  "best_metric": 0.09392726059392727,
  "best_model_checkpoint": "fine-w2v2base-bs16-ep100-lr2e-05-linguistic-rmsnorm-focal_ctc_a0.25_g0.5-0.05_10_0.004_40/checkpoint-3000",
  "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": 8.301886792452831e-07,
      "loss": 522.579,
      "step": 50
    },
    {
      "epoch": 0.94,
      "eval_loss": 251.04580688476562,
      "eval_runtime": 2.959,
      "eval_samples_per_second": 242.309,
      "eval_steps_per_second": 7.773,
      "eval_wer": 14.963880547213881,
      "step": 50
    },
    {
      "epoch": 1.89,
      "learning_rate": 2.6415094339622644e-06,
      "loss": 317.1515,
      "step": 100
    },
    {
      "epoch": 1.89,
      "eval_loss": 84.86875915527344,
      "eval_runtime": 2.2882,
      "eval_samples_per_second": 313.342,
      "eval_steps_per_second": 10.051,
      "eval_wer": 0.9970804137470805,
      "step": 100
    },
    {
      "epoch": 2.83,
      "learning_rate": 4.528301886792453e-06,
      "loss": 57.5912,
      "step": 150
    },
    {
      "epoch": 2.83,
      "eval_loss": 24.78022003173828,
      "eval_runtime": 2.3009,
      "eval_samples_per_second": 311.621,
      "eval_steps_per_second": 9.996,
      "eval_wer": 1.0,
      "step": 150
    },
    {
      "epoch": 3.77,
      "learning_rate": 6.415094339622642e-06,
      "loss": 28.4209,
      "step": 200
    },
    {
      "epoch": 3.77,
      "eval_loss": 21.716365814208984,
      "eval_runtime": 2.2928,
      "eval_samples_per_second": 312.721,
      "eval_steps_per_second": 10.031,
      "eval_wer": 1.0,
      "step": 200
    },
    {
      "epoch": 4.72,
      "learning_rate": 8.301886792452832e-06,
      "loss": 27.1215,
      "step": 250
    },
    {
      "epoch": 4.72,
      "eval_loss": 21.151853561401367,
      "eval_runtime": 2.2677,
      "eval_samples_per_second": 316.176,
      "eval_steps_per_second": 10.142,
      "eval_wer": 1.0,
      "step": 250
    },
    {
      "epoch": 5.66,
      "learning_rate": 1.018867924528302e-05,
      "loss": 26.1663,
      "step": 300
    },
    {
      "epoch": 5.66,
      "eval_loss": 20.574935913085938,
      "eval_runtime": 2.409,
      "eval_samples_per_second": 297.634,
      "eval_steps_per_second": 9.548,
      "eval_wer": 1.0,
      "step": 300
    },
    {
      "epoch": 6.6,
      "learning_rate": 1.2075471698113209e-05,
      "loss": 25.4374,
      "step": 350
    },
    {
      "epoch": 6.6,
      "eval_loss": 20.153608322143555,
      "eval_runtime": 2.3039,
      "eval_samples_per_second": 311.217,
      "eval_steps_per_second": 9.983,
      "eval_wer": 1.0,
      "step": 350
    },
    {
      "epoch": 7.55,
      "learning_rate": 1.3962264150943397e-05,
      "loss": 24.5548,
      "step": 400
    },
    {
      "epoch": 7.55,
      "eval_loss": 19.66973304748535,
      "eval_runtime": 2.3227,
      "eval_samples_per_second": 308.693,
      "eval_steps_per_second": 9.902,
      "eval_wer": 1.0,
      "step": 400
    },
    {
      "epoch": 8.49,
      "learning_rate": 1.5849056603773586e-05,
      "loss": 24.1548,
      "step": 450
    },
    {
      "epoch": 8.49,
      "eval_loss": 19.558759689331055,
      "eval_runtime": 2.2394,
      "eval_samples_per_second": 320.177,
      "eval_steps_per_second": 10.271,
      "eval_wer": 1.0,
      "step": 450
    },
    {
      "epoch": 9.43,
      "learning_rate": 1.7735849056603774e-05,
      "loss": 24.4262,
      "step": 500
    },
    {
      "epoch": 9.43,
      "eval_loss": 19.487014770507812,
      "eval_runtime": 2.2407,
      "eval_samples_per_second": 319.996,
      "eval_steps_per_second": 10.265,
      "eval_wer": 1.0,
      "step": 500
    },
    {
      "epoch": 10.38,
      "learning_rate": 1.9622641509433963e-05,
      "loss": 24.0949,
      "step": 550
    },
    {
      "epoch": 10.38,
      "eval_loss": 19.597867965698242,
      "eval_runtime": 2.4442,
      "eval_samples_per_second": 293.346,
      "eval_steps_per_second": 9.41,
      "eval_wer": 1.0,
      "step": 550
    },
    {
      "epoch": 11.32,
      "learning_rate": 1.9996530012933285e-05,
      "loss": 24.1762,
      "step": 600
    },
    {
      "epoch": 11.32,
      "eval_loss": 20.214019775390625,
      "eval_runtime": 2.3713,
      "eval_samples_per_second": 302.362,
      "eval_steps_per_second": 9.699,
      "eval_wer": 1.0,
      "step": 600
    },
    {
      "epoch": 12.26,
      "learning_rate": 1.9982437317643218e-05,
      "loss": 23.2554,
      "step": 650
    },
    {
      "epoch": 12.26,
      "eval_loss": 20.086496353149414,
      "eval_runtime": 2.3649,
      "eval_samples_per_second": 303.179,
      "eval_steps_per_second": 9.725,
      "eval_wer": 1.0,
      "step": 650
    },
    {
      "epoch": 13.21,
      "learning_rate": 1.9957520309390786e-05,
      "loss": 22.7304,
      "step": 700
    },
    {
      "epoch": 13.21,
      "eval_loss": 19.462364196777344,
      "eval_runtime": 2.3545,
      "eval_samples_per_second": 304.518,
      "eval_steps_per_second": 9.768,
      "eval_wer": 0.9999165832499166,
      "step": 700
    },
    {
      "epoch": 14.15,
      "learning_rate": 1.9921806006577102e-05,
      "loss": 22.0028,
      "step": 750
    },
    {
      "epoch": 14.15,
      "eval_loss": 17.890724182128906,
      "eval_runtime": 2.3072,
      "eval_samples_per_second": 310.77,
      "eval_steps_per_second": 9.969,
      "eval_wer": 0.9990824157490824,
      "step": 750
    },
    {
      "epoch": 15.09,
      "learning_rate": 1.9875333135495e-05,
      "loss": 20.0064,
      "step": 800
    },
    {
      "epoch": 15.09,
      "eval_loss": 11.938626289367676,
      "eval_runtime": 2.3126,
      "eval_samples_per_second": 310.046,
      "eval_steps_per_second": 9.946,
      "eval_wer": 0.7019519519519519,
      "step": 800
    },
    {
      "epoch": 16.04,
      "learning_rate": 1.9818152088336786e-05,
      "loss": 12.6884,
      "step": 850
    },
    {
      "epoch": 16.04,
      "eval_loss": 5.536014080047607,
      "eval_runtime": 2.2922,
      "eval_samples_per_second": 312.806,
      "eval_steps_per_second": 10.034,
      "eval_wer": 0.3160660660660661,
      "step": 850
    },
    {
      "epoch": 16.98,
      "learning_rate": 1.9750324868552133e-05,
      "loss": 7.2843,
      "step": 900
    },
    {
      "epoch": 16.98,
      "eval_loss": 3.463259696960449,
      "eval_runtime": 2.2577,
      "eval_samples_per_second": 317.586,
      "eval_steps_per_second": 10.188,
      "eval_wer": 0.21054387721054388,
      "step": 900
    },
    {
      "epoch": 17.92,
      "learning_rate": 1.9671925023615572e-05,
      "loss": 5.2335,
      "step": 950
    },
    {
      "epoch": 17.92,
      "eval_loss": 2.6680145263671875,
      "eval_runtime": 2.2729,
      "eval_samples_per_second": 315.45,
      "eval_steps_per_second": 10.119,
      "eval_wer": 0.17258925592258925,
      "step": 950
    },
    {
      "epoch": 18.87,
      "learning_rate": 1.9583037565276314e-05,
      "loss": 4.2601,
      "step": 1000
    },
    {
      "epoch": 18.87,
      "eval_loss": 2.1859302520751953,
      "eval_runtime": 2.3197,
      "eval_samples_per_second": 309.093,
      "eval_steps_per_second": 9.915,
      "eval_wer": 0.1504838171504838,
      "step": 1000
    },
    {
      "epoch": 19.81,
      "learning_rate": 1.948375887737699e-05,
      "loss": 3.6512,
      "step": 1050
    },
    {
      "epoch": 19.81,
      "eval_loss": 1.9664338827133179,
      "eval_runtime": 2.2432,
      "eval_samples_per_second": 319.626,
      "eval_steps_per_second": 10.253,
      "eval_wer": 0.14714714714714713,
      "step": 1050
    },
    {
      "epoch": 20.75,
      "learning_rate": 1.9374196611341212e-05,
      "loss": 3.2164,
      "step": 1100
    },
    {
      "epoch": 20.75,
      "eval_loss": 1.7849069833755493,
      "eval_runtime": 2.2582,
      "eval_samples_per_second": 317.504,
      "eval_steps_per_second": 10.185,
      "eval_wer": 0.13505171838505173,
      "step": 1100
    },
    {
      "epoch": 21.7,
      "learning_rate": 1.9254469569443274e-05,
      "loss": 3.0286,
      "step": 1150
    },
    {
      "epoch": 21.7,
      "eval_loss": 1.6424585580825806,
      "eval_runtime": 2.296,
      "eval_samples_per_second": 312.288,
      "eval_steps_per_second": 10.018,
      "eval_wer": 0.13129796463129798,
      "step": 1150
    },
    {
      "epoch": 22.64,
      "learning_rate": 1.9124707575986642e-05,
      "loss": 2.776,
      "step": 1200
    },
    {
      "epoch": 22.64,
      "eval_loss": 1.5508594512939453,
      "eval_runtime": 2.314,
      "eval_samples_per_second": 309.857,
      "eval_steps_per_second": 9.94,
      "eval_wer": 0.13229896563229895,
      "step": 1200
    },
    {
      "epoch": 23.58,
      "learning_rate": 1.89850513365308e-05,
      "loss": 2.5805,
      "step": 1250
    },
    {
      "epoch": 23.58,
      "eval_loss": 1.5048248767852783,
      "eval_runtime": 2.3299,
      "eval_samples_per_second": 307.745,
      "eval_steps_per_second": 9.872,
      "eval_wer": 0.12812812812812813,
      "step": 1250
    },
    {
      "epoch": 24.53,
      "learning_rate": 1.883565228531919e-05,
      "loss": 2.372,
      "step": 1300
    },
    {
      "epoch": 24.53,
      "eval_loss": 1.4449571371078491,
      "eval_runtime": 2.2519,
      "eval_samples_per_second": 318.396,
      "eval_steps_per_second": 10.214,
      "eval_wer": 0.11686686686686687,
      "step": 1300
    },
    {
      "epoch": 25.47,
      "learning_rate": 1.86766724210737e-05,
      "loss": 2.3566,
      "step": 1350
    },
    {
      "epoch": 25.47,
      "eval_loss": 1.3800303936004639,
      "eval_runtime": 2.314,
      "eval_samples_per_second": 309.859,
      "eval_steps_per_second": 9.94,
      "eval_wer": 0.11361361361361362,
      "step": 1350
    },
    {
      "epoch": 26.42,
      "learning_rate": 1.8508284131333604e-05,
      "loss": 2.137,
      "step": 1400
    },
    {
      "epoch": 26.42,
      "eval_loss": 1.353422999382019,
      "eval_runtime": 2.3077,
      "eval_samples_per_second": 310.7,
      "eval_steps_per_second": 9.967,
      "eval_wer": 0.11644978311644978,
      "step": 1400
    },
    {
      "epoch": 27.36,
      "learning_rate": 1.8330670005529657e-05,
      "loss": 2.1112,
      "step": 1450
    },
    {
      "epoch": 27.36,
      "eval_loss": 1.3263455629348755,
      "eval_runtime": 2.3816,
      "eval_samples_per_second": 301.058,
      "eval_steps_per_second": 9.657,
      "eval_wer": 0.11319652986319653,
      "step": 1450
    },
    {
      "epoch": 28.3,
      "learning_rate": 1.814402263699584e-05,
      "loss": 1.9889,
      "step": 1500
    },
    {
      "epoch": 28.3,
      "eval_loss": 1.3026772737503052,
      "eval_runtime": 2.3127,
      "eval_samples_per_second": 310.023,
      "eval_steps_per_second": 9.945,
      "eval_wer": 0.1091091091091091,
      "step": 1500
    },
    {
      "epoch": 29.25,
      "learning_rate": 1.7952539165550863e-05,
      "loss": 1.9183,
      "step": 1550
    },
    {
      "epoch": 29.25,
      "eval_loss": 1.2998329401016235,
      "eval_runtime": 2.2784,
      "eval_samples_per_second": 314.692,
      "eval_steps_per_second": 10.095,
      "eval_wer": 0.11169502836169502,
      "step": 1550
    },
    {
      "epoch": 30.19,
      "learning_rate": 1.7748612292093336e-05,
      "loss": 1.8744,
      "step": 1600
    },
    {
      "epoch": 30.19,
      "eval_loss": 1.2636977434158325,
      "eval_runtime": 2.3196,
      "eval_samples_per_second": 309.103,
      "eval_steps_per_second": 9.915,
      "eval_wer": 0.10927594260927594,
      "step": 1600
    },
    {
      "epoch": 31.13,
      "learning_rate": 1.753628332185275e-05,
      "loss": 1.75,
      "step": 1650
    },
    {
      "epoch": 31.13,
      "eval_loss": 1.271153450012207,
      "eval_runtime": 2.2663,
      "eval_samples_per_second": 316.378,
      "eval_steps_per_second": 10.149,
      "eval_wer": 0.10593927260593927,
      "step": 1650
    },
    {
      "epoch": 32.08,
      "learning_rate": 1.731578249070756e-05,
      "loss": 1.7865,
      "step": 1700
    },
    {
      "epoch": 32.08,
      "eval_loss": 1.2367671728134155,
      "eval_runtime": 2.2605,
      "eval_samples_per_second": 317.191,
      "eval_steps_per_second": 10.175,
      "eval_wer": 0.10919252585919252,
      "step": 1700
    },
    {
      "epoch": 33.02,
      "learning_rate": 1.7087348895576564e-05,
      "loss": 1.6976,
      "step": 1750
    },
    {
      "epoch": 33.02,
      "eval_loss": 1.2081226110458374,
      "eval_runtime": 2.2932,
      "eval_samples_per_second": 312.658,
      "eval_steps_per_second": 10.029,
      "eval_wer": 0.10385385385385386,
      "step": 1750
    },
    {
      "epoch": 33.96,
      "learning_rate": 1.68512302351576e-05,
      "loss": 1.6891,
      "step": 1800
    },
    {
      "epoch": 33.96,
      "eval_loss": 1.2145798206329346,
      "eval_runtime": 2.2483,
      "eval_samples_per_second": 318.902,
      "eval_steps_per_second": 10.23,
      "eval_wer": 0.10643977310643978,
      "step": 1800
    },
    {
      "epoch": 34.91,
      "learning_rate": 1.6607682541338998e-05,
      "loss": 1.5919,
      "step": 1850
    },
    {
      "epoch": 34.91,
      "eval_loss": 1.208232045173645,
      "eval_runtime": 2.2769,
      "eval_samples_per_second": 314.903,
      "eval_steps_per_second": 10.101,
      "eval_wer": 0.10802469135802469,
      "step": 1850
    },
    {
      "epoch": 35.85,
      "learning_rate": 1.6356969901575094e-05,
      "loss": 1.5751,
      "step": 1900
    },
    {
      "epoch": 35.85,
      "eval_loss": 1.2008248567581177,
      "eval_runtime": 2.2525,
      "eval_samples_per_second": 318.319,
      "eval_steps_per_second": 10.211,
      "eval_wer": 0.10326993660326994,
      "step": 1900
    },
    {
      "epoch": 36.79,
      "learning_rate": 1.6099364172526732e-05,
      "loss": 1.5628,
      "step": 1950
    },
    {
      "epoch": 36.79,
      "eval_loss": 1.1641323566436768,
      "eval_runtime": 2.2858,
      "eval_samples_per_second": 313.681,
      "eval_steps_per_second": 10.062,
      "eval_wer": 0.10243576910243576,
      "step": 1950
    },
    {
      "epoch": 37.74,
      "learning_rate": 1.583514468527744e-05,
      "loss": 1.4812,
      "step": 2000
    },
    {
      "epoch": 37.74,
      "eval_loss": 1.2022244930267334,
      "eval_runtime": 2.3066,
      "eval_samples_per_second": 310.852,
      "eval_steps_per_second": 9.972,
      "eval_wer": 0.10535535535535535,
      "step": 2000
    },
    {
      "epoch": 38.68,
      "learning_rate": 1.5564597942444743e-05,
      "loss": 1.4784,
      "step": 2050
    },
    {
      "epoch": 38.68,
      "eval_loss": 1.1667057275772095,
      "eval_runtime": 2.336,
      "eval_samples_per_second": 306.934,
      "eval_steps_per_second": 9.846,
      "eval_wer": 0.10251918585251918,
      "step": 2050
    },
    {
      "epoch": 39.62,
      "learning_rate": 1.5288017307515142e-05,
      "loss": 1.4142,
      "step": 2100
    },
    {
      "epoch": 39.62,
      "eval_loss": 1.1611201763153076,
      "eval_runtime": 2.4404,
      "eval_samples_per_second": 293.809,
      "eval_steps_per_second": 9.425,
      "eval_wer": 0.10535535535535535,
      "step": 2100
    },
    {
      "epoch": 40.57,
      "learning_rate": 1.500570268673965e-05,
      "loss": 1.3841,
      "step": 2150
    },
    {
      "epoch": 40.57,
      "eval_loss": 1.125167727470398,
      "eval_runtime": 2.4422,
      "eval_samples_per_second": 293.593,
      "eval_steps_per_second": 9.418,
      "eval_wer": 0.09793126459793126,
      "step": 2150
    },
    {
      "epoch": 41.51,
      "learning_rate": 1.4717960203934704e-05,
      "loss": 1.3636,
      "step": 2200
    },
    {
      "epoch": 41.51,
      "eval_loss": 1.1582497358322144,
      "eval_runtime": 2.3347,
      "eval_samples_per_second": 307.1,
      "eval_steps_per_second": 9.851,
      "eval_wer": 0.10218551885218552,
      "step": 2200
    },
    {
      "epoch": 42.45,
      "learning_rate": 1.4425101868541228e-05,
      "loss": 1.3526,
      "step": 2250
    },
    {
      "epoch": 42.45,
      "eval_loss": 1.1616238355636597,
      "eval_runtime": 2.2766,
      "eval_samples_per_second": 314.937,
      "eval_steps_per_second": 10.103,
      "eval_wer": 0.10802469135802469,
      "step": 2250
    },
    {
      "epoch": 43.4,
      "learning_rate": 1.412744523730163e-05,
      "loss": 1.2923,
      "step": 2300
    },
    {
      "epoch": 43.4,
      "eval_loss": 1.1713889837265015,
      "eval_runtime": 2.275,
      "eval_samples_per_second": 315.169,
      "eval_steps_per_second": 10.11,
      "eval_wer": 0.10452118785452119,
      "step": 2300
    },
    {
      "epoch": 44.34,
      "learning_rate": 1.3825313069921713e-05,
      "loss": 1.2576,
      "step": 2350
    },
    {
      "epoch": 44.34,
      "eval_loss": 1.1561468839645386,
      "eval_runtime": 2.2633,
      "eval_samples_per_second": 316.792,
      "eval_steps_per_second": 10.162,
      "eval_wer": 0.10352018685352019,
      "step": 2350
    },
    {
      "epoch": 45.28,
      "learning_rate": 1.3519032979090816e-05,
      "loss": 1.2791,
      "step": 2400
    },
    {
      "epoch": 45.28,
      "eval_loss": 1.1193358898162842,
      "eval_runtime": 2.3275,
      "eval_samples_per_second": 308.053,
      "eval_steps_per_second": 9.882,
      "eval_wer": 0.1006006006006006,
      "step": 2400
    },
    {
      "epoch": 46.23,
      "learning_rate": 1.3208937075239663e-05,
      "loss": 1.2104,
      "step": 2450
    },
    {
      "epoch": 46.23,
      "eval_loss": 1.1346070766448975,
      "eval_runtime": 2.3774,
      "eval_samples_per_second": 301.59,
      "eval_steps_per_second": 9.674,
      "eval_wer": 0.10260260260260261,
      "step": 2450
    },
    {
      "epoch": 47.17,
      "learning_rate": 1.2901665022139796e-05,
      "loss": 1.1839,
      "step": 2500
    },
    {
      "epoch": 47.17,
      "eval_loss": 1.112608790397644,
      "eval_runtime": 2.2603,
      "eval_samples_per_second": 317.217,
      "eval_steps_per_second": 10.176,
      "eval_wer": 0.10093426760093427,
      "step": 2500
    },
    {
      "epoch": 48.11,
      "learning_rate": 1.2585009445463867e-05,
      "loss": 1.1314,
      "step": 2550
    },
    {
      "epoch": 48.11,
      "eval_loss": 1.1136256456375122,
      "eval_runtime": 2.3442,
      "eval_samples_per_second": 305.863,
      "eval_steps_per_second": 9.811,
      "eval_wer": 0.0995995995995996,
      "step": 2550
    },
    {
      "epoch": 49.06,
      "learning_rate": 1.2265550850811663e-05,
      "loss": 1.1772,
      "step": 2600
    },
    {
      "epoch": 49.06,
      "eval_loss": 1.136879563331604,
      "eval_runtime": 2.4082,
      "eval_samples_per_second": 297.727,
      "eval_steps_per_second": 9.551,
      "eval_wer": 0.10293626960293627,
      "step": 2600
    },
    {
      "epoch": 50.0,
      "learning_rate": 1.1943635638535827e-05,
      "loss": 1.1137,
      "step": 2650
    },
    {
      "epoch": 50.0,
      "eval_loss": 1.1156716346740723,
      "eval_runtime": 2.2442,
      "eval_samples_per_second": 319.494,
      "eval_steps_per_second": 10.249,
      "eval_wer": 0.10118451785118451,
      "step": 2650
    },
    {
      "epoch": 50.94,
      "learning_rate": 1.1619612872787144e-05,
      "loss": 1.1125,
      "step": 2700
    },
    {
      "epoch": 50.94,
      "eval_loss": 1.1241499185562134,
      "eval_runtime": 2.2664,
      "eval_samples_per_second": 316.356,
      "eval_steps_per_second": 10.148,
      "eval_wer": 0.10151818485151819,
      "step": 2700
    },
    {
      "epoch": 51.89,
      "learning_rate": 1.1293833903011819e-05,
      "loss": 1.1536,
      "step": 2750
    },
    {
      "epoch": 51.89,
      "eval_loss": 1.1277304887771606,
      "eval_runtime": 2.3054,
      "eval_samples_per_second": 311.008,
      "eval_steps_per_second": 9.977,
      "eval_wer": 0.10118451785118451,
      "step": 2750
    },
    {
      "epoch": 52.83,
      "learning_rate": 1.0966651982970757e-05,
      "loss": 1.0589,
      "step": 2800
    },
    {
      "epoch": 52.83,
      "eval_loss": 1.1413406133651733,
      "eval_runtime": 2.3376,
      "eval_samples_per_second": 306.721,
      "eval_steps_per_second": 9.839,
      "eval_wer": 0.11419753086419752,
      "step": 2800
    },
    {
      "epoch": 53.77,
      "learning_rate": 1.0638421887693887e-05,
      "loss": 1.1234,
      "step": 2850
    },
    {
      "epoch": 53.77,
      "eval_loss": 1.1188369989395142,
      "eval_runtime": 2.4174,
      "eval_samples_per_second": 296.596,
      "eval_steps_per_second": 9.514,
      "eval_wer": 0.10335335335335336,
      "step": 2850
    },
    {
      "epoch": 54.72,
      "learning_rate": 1.0309499528784948e-05,
      "loss": 1.1047,
      "step": 2900
    },
    {
      "epoch": 54.72,
      "eval_loss": 1.1185845136642456,
      "eval_runtime": 2.2736,
      "eval_samples_per_second": 315.361,
      "eval_steps_per_second": 10.116,
      "eval_wer": 0.10677344010677343,
      "step": 2900
    },
    {
      "epoch": 55.66,
      "learning_rate": 9.980241568493834e-06,
      "loss": 0.9979,
      "step": 2950
    },
    {
      "epoch": 55.66,
      "eval_loss": 1.1078546047210693,
      "eval_runtime": 2.3692,
      "eval_samples_per_second": 302.631,
      "eval_steps_per_second": 9.708,
      "eval_wer": 0.10068401735068402,
      "step": 2950
    },
    {
      "epoch": 56.6,
      "learning_rate": 9.651005032974994e-06,
      "loss": 0.9788,
      "step": 3000
    },
    {
      "epoch": 56.6,
      "eval_loss": 1.0918303728103638,
      "eval_runtime": 2.2466,
      "eval_samples_per_second": 319.145,
      "eval_steps_per_second": 10.238,
      "eval_wer": 0.09392726059392727,
      "step": 3000
    },
    {
      "epoch": 57.55,
      "learning_rate": 9.322146925151226e-06,
      "loss": 1.009,
      "step": 3050
    },
    {
      "epoch": 57.55,
      "eval_loss": 1.117212176322937,
      "eval_runtime": 2.3635,
      "eval_samples_per_second": 303.36,
      "eval_steps_per_second": 9.731,
      "eval_wer": 0.10235235235235235,
      "step": 3050
    },
    {
      "epoch": 58.49,
      "learning_rate": 8.994023837602694e-06,
      "loss": 0.9942,
      "step": 3100
    },
    {
      "epoch": 58.49,
      "eval_loss": 1.1138882637023926,
      "eval_runtime": 2.2403,
      "eval_samples_per_second": 320.042,
      "eval_steps_per_second": 10.266,
      "eval_wer": 0.09901568234901569,
      "step": 3100
    },
    {
      "epoch": 59.43,
      "learning_rate": 8.666991565900827e-06,
      "loss": 0.9602,
      "step": 3150
    },
    {
      "epoch": 59.43,
      "eval_loss": 1.1062616109848022,
      "eval_runtime": 2.276,
      "eval_samples_per_second": 315.031,
      "eval_steps_per_second": 10.106,
      "eval_wer": 0.10168501835168502,
      "step": 3150
    },
    {
      "epoch": 60.38,
      "learning_rate": 8.341404722806525e-06,
      "loss": 0.9813,
      "step": 3200
    },
    {
      "epoch": 60.38,
      "eval_loss": 1.1151028871536255,
      "eval_runtime": 2.2474,
      "eval_samples_per_second": 319.038,
      "eval_steps_per_second": 10.234,
      "eval_wer": 0.10468802135468802,
      "step": 3200
    },
    {
      "epoch": 61.32,
      "learning_rate": 8.017616353750874e-06,
      "loss": 0.9112,
      "step": 3250
    },
    {
      "epoch": 61.32,
      "eval_loss": 1.0930062532424927,
      "eval_runtime": 2.3667,
      "eval_samples_per_second": 302.948,
      "eval_steps_per_second": 9.718,
      "eval_wer": 0.09701368034701369,
      "step": 3250
    },
    {
      "epoch": 62.26,
      "learning_rate": 7.695977554015387e-06,
      "loss": 0.9705,
      "step": 3300
    },
    {
      "epoch": 62.26,
      "eval_loss": 1.098963737487793,
      "eval_runtime": 2.2883,
      "eval_samples_per_second": 313.338,
      "eval_steps_per_second": 10.051,
      "eval_wer": 0.09926593259926593,
      "step": 3300
    },
    {
      "epoch": 63.21,
      "learning_rate": 7.376837088026863e-06,
      "loss": 0.8753,
      "step": 3350
    },
    {
      "epoch": 63.21,
      "eval_loss": 1.105305790901184,
      "eval_runtime": 2.2693,
      "eval_samples_per_second": 315.96,
      "eval_steps_per_second": 10.135,
      "eval_wer": 0.10385385385385386,
      "step": 3350
    },
    {
      "epoch": 64.15,
      "learning_rate": 7.0605410111796855e-06,
      "loss": 0.9259,
      "step": 3400
    },
    {
      "epoch": 64.15,
      "eval_loss": 1.0978156328201294,
      "eval_runtime": 2.2619,
      "eval_samples_per_second": 316.996,
      "eval_steps_per_second": 10.169,
      "eval_wer": 0.09843176509843177,
      "step": 3400
    },
    {
      "epoch": 65.09,
      "learning_rate": 6.747432294595591e-06,
      "loss": 0.8877,
      "step": 3450
    },
    {
      "epoch": 65.09,
      "eval_loss": 1.1047441959381104,
      "eval_runtime": 2.4284,
      "eval_samples_per_second": 295.252,
      "eval_steps_per_second": 9.471,
      "eval_wer": 0.09868201534868201,
      "step": 3450
    },
    {
      "epoch": 66.04,
      "learning_rate": 6.437850453227872e-06,
      "loss": 0.9111,
      "step": 3500
    },
    {
      "epoch": 66.04,
      "eval_loss": 1.093747615814209,
      "eval_runtime": 2.3087,
      "eval_samples_per_second": 310.56,
      "eval_steps_per_second": 9.962,
      "eval_wer": 0.10093426760093427,
      "step": 3500
    },
    {
      "epoch": 66.98,
      "learning_rate": 6.132131177713165e-06,
      "loss": 0.9103,
      "step": 3550
    },
    {
      "epoch": 66.98,
      "eval_loss": 1.0963308811187744,
      "eval_runtime": 2.4196,
      "eval_samples_per_second": 296.334,
      "eval_steps_per_second": 9.506,
      "eval_wer": 0.09984984984984985,
      "step": 3550
    },
    {
      "epoch": 67.92,
      "learning_rate": 5.830605970370142e-06,
      "loss": 0.9031,
      "step": 3600
    },
    {
      "epoch": 67.92,
      "eval_loss": 1.096879243850708,
      "eval_runtime": 2.371,
      "eval_samples_per_second": 302.403,
      "eval_steps_per_second": 9.701,
      "eval_wer": 0.10243576910243576,
      "step": 3600
    },
    {
      "epoch": 68.87,
      "learning_rate": 5.533601785739714e-06,
      "loss": 0.876,
      "step": 3650
    },
    {
      "epoch": 68.87,
      "eval_loss": 1.092012882232666,
      "eval_runtime": 2.2592,
      "eval_samples_per_second": 317.366,
      "eval_steps_per_second": 10.181,
      "eval_wer": 0.09642976309642977,
      "step": 3650
    },
    {
      "epoch": 69.81,
      "learning_rate": 5.24144067605655e-06,
      "loss": 0.8722,
      "step": 3700
    },
    {
      "epoch": 69.81,
      "eval_loss": 1.0867727994918823,
      "eval_runtime": 2.3239,
      "eval_samples_per_second": 308.535,
      "eval_steps_per_second": 9.897,
      "eval_wer": 0.09584584584584585,
      "step": 3700
    },
    {
      "epoch": 70.75,
      "learning_rate": 4.9544394420363395e-06,
      "loss": 0.8751,
      "step": 3750
    },
    {
      "epoch": 70.75,
      "eval_loss": 1.087986707687378,
      "eval_runtime": 2.4282,
      "eval_samples_per_second": 295.277,
      "eval_steps_per_second": 9.472,
      "eval_wer": 0.0965965965965966,
      "step": 3750
    },
    {
      "epoch": 71.7,
      "learning_rate": 4.672909289357498e-06,
      "loss": 0.8816,
      "step": 3800
    },
    {
      "epoch": 71.7,
      "eval_loss": 1.0878993272781372,
      "eval_runtime": 2.265,
      "eval_samples_per_second": 316.559,
      "eval_steps_per_second": 10.155,
      "eval_wer": 0.09743076409743076,
      "step": 3800
    },
    {
      "epoch": 72.64,
      "learning_rate": 4.397155491209727e-06,
      "loss": 0.8488,
      "step": 3850
    },
    {
      "epoch": 72.64,
      "eval_loss": 1.0897611379623413,
      "eval_runtime": 2.2509,
      "eval_samples_per_second": 318.535,
      "eval_steps_per_second": 10.218,
      "eval_wer": 0.09743076409743076,
      "step": 3850
    },
    {
      "epoch": 73.58,
      "learning_rate": 4.127477057275398e-06,
      "loss": 0.8327,
      "step": 3900
    },
    {
      "epoch": 73.58,
      "eval_loss": 1.0847886800765991,
      "eval_runtime": 2.4261,
      "eval_samples_per_second": 295.54,
      "eval_steps_per_second": 9.48,
      "eval_wer": 0.09776443109776443,
      "step": 3900
    },
    {
      "epoch": 74.53,
      "learning_rate": 3.864166409502706e-06,
      "loss": 0.7818,
      "step": 3950
    },
    {
      "epoch": 74.53,
      "eval_loss": 1.0878225564956665,
      "eval_runtime": 2.4194,
      "eval_samples_per_second": 296.349,
      "eval_steps_per_second": 9.506,
      "eval_wer": 0.09567901234567901,
      "step": 3950
    },
    {
      "epoch": 75.47,
      "learning_rate": 3.607509065022101e-06,
      "loss": 0.8569,
      "step": 4000
    },
    {
      "epoch": 75.47,
      "eval_loss": 1.0838165283203125,
      "eval_runtime": 2.2558,
      "eval_samples_per_second": 317.848,
      "eval_steps_per_second": 10.196,
      "eval_wer": 0.09968301634968302,
      "step": 4000
    },
    {
      "epoch": 76.42,
      "learning_rate": 3.3577833265498728e-06,
      "loss": 0.8078,
      "step": 4050
    },
    {
      "epoch": 76.42,
      "eval_loss": 1.0724711418151855,
      "eval_runtime": 2.4159,
      "eval_samples_per_second": 296.781,
      "eval_steps_per_second": 9.52,
      "eval_wer": 0.09826493159826494,
      "step": 4050
    },
    {
      "epoch": 77.36,
      "learning_rate": 3.115259980614602e-06,
      "loss": 0.8557,
      "step": 4100
    },
    {
      "epoch": 77.36,
      "eval_loss": 1.0776066780090332,
      "eval_runtime": 2.4314,
      "eval_samples_per_second": 294.887,
      "eval_steps_per_second": 9.459,
      "eval_wer": 0.10001668335001668,
      "step": 4100
    },
    {
      "epoch": 78.3,
      "learning_rate": 2.880202003933645e-06,
      "loss": 0.8361,
      "step": 4150
    },
    {
      "epoch": 78.3,
      "eval_loss": 1.085669994354248,
      "eval_runtime": 2.2432,
      "eval_samples_per_second": 319.637,
      "eval_steps_per_second": 10.253,
      "eval_wer": 0.09776443109776443,
      "step": 4150
    },
    {
      "epoch": 79.25,
      "learning_rate": 2.652864278258126e-06,
      "loss": 0.7911,
      "step": 4200
    },
    {
      "epoch": 79.25,
      "eval_loss": 1.081592321395874,
      "eval_runtime": 2.2789,
      "eval_samples_per_second": 314.631,
      "eval_steps_per_second": 10.093,
      "eval_wer": 0.09526192859526193,
      "step": 4200
    },
    {
      "epoch": 80.19,
      "learning_rate": 2.433493313995524e-06,
      "loss": 0.8146,
      "step": 4250
    },
    {
      "epoch": 80.19,
      "eval_loss": 1.0815632343292236,
      "eval_runtime": 2.274,
      "eval_samples_per_second": 315.3,
      "eval_steps_per_second": 10.114,
      "eval_wer": 0.09693026359693026,
      "step": 4250
    },
    {
      "epoch": 81.13,
      "learning_rate": 2.2223269829096593e-06,
      "loss": 0.8237,
      "step": 4300
    },
    {
      "epoch": 81.13,
      "eval_loss": 1.0928122997283936,
      "eval_runtime": 2.2839,
      "eval_samples_per_second": 313.931,
      "eval_steps_per_second": 10.07,
      "eval_wer": 0.10051718385051718,
      "step": 4300
    },
    {
      "epoch": 82.08,
      "learning_rate": 2.0195942601878703e-06,
      "loss": 0.7944,
      "step": 4350
    },
    {
      "epoch": 82.08,
      "eval_loss": 1.0918399095535278,
      "eval_runtime": 2.3844,
      "eval_samples_per_second": 300.701,
      "eval_steps_per_second": 9.646,
      "eval_wer": 0.09651317984651318,
      "step": 4350
    },
    {
      "epoch": 83.02,
      "learning_rate": 1.8255149761550128e-06,
      "loss": 0.8108,
      "step": 4400
    },
    {
      "epoch": 83.02,
      "eval_loss": 1.094603180885315,
      "eval_runtime": 2.3921,
      "eval_samples_per_second": 299.741,
      "eval_steps_per_second": 9.615,
      "eval_wer": 0.09684684684684684,
      "step": 4400
    },
    {
      "epoch": 83.96,
      "learning_rate": 1.6402995779036146e-06,
      "loss": 0.7892,
      "step": 4450
    },
    {
      "epoch": 83.96,
      "eval_loss": 1.0920747518539429,
      "eval_runtime": 2.2573,
      "eval_samples_per_second": 317.631,
      "eval_steps_per_second": 10.189,
      "eval_wer": 0.09684684684684684,
      "step": 4450
    },
    {
      "epoch": 84.91,
      "learning_rate": 1.4641489010985954e-06,
      "loss": 0.8261,
      "step": 4500
    },
    {
      "epoch": 84.91,
      "eval_loss": 1.086715579032898,
      "eval_runtime": 2.2828,
      "eval_samples_per_second": 314.092,
      "eval_steps_per_second": 10.075,
      "eval_wer": 0.09751418084751418,
      "step": 4500
    },
    {
      "epoch": 85.85,
      "learning_rate": 1.3004999577835786e-06,
      "loss": 0.7909,
      "step": 4550
    },
    {
      "epoch": 85.85,
      "eval_loss": 1.0858174562454224,
      "eval_runtime": 2.2891,
      "eval_samples_per_second": 313.219,
      "eval_steps_per_second": 10.047,
      "eval_wer": 0.09642976309642977,
      "step": 4550
    },
    {
      "epoch": 86.79,
      "learning_rate": 1.1428512653500146e-06,
      "loss": 0.804,
      "step": 4600
    },
    {
      "epoch": 86.79,
      "eval_loss": 1.0831711292266846,
      "eval_runtime": 2.2911,
      "eval_samples_per_second": 312.953,
      "eval_steps_per_second": 10.039,
      "eval_wer": 0.0965965965965966,
      "step": 4600
    },
    {
      "epoch": 87.74,
      "learning_rate": 9.94806695317354e-07,
      "loss": 0.7981,
      "step": 4650
    },
    {
      "epoch": 87.74,
      "eval_loss": 1.0887707471847534,
      "eval_runtime": 2.2547,
      "eval_samples_per_second": 318.005,
      "eval_steps_per_second": 10.201,
      "eval_wer": 0.09843176509843177,
      "step": 4650
    },
    {
      "epoch": 88.68,
      "learning_rate": 8.56526777695198e-07,
      "loss": 0.7975,
      "step": 4700
    },
    {
      "epoch": 88.68,
      "eval_loss": 1.0889778137207031,
      "eval_runtime": 2.4661,
      "eval_samples_per_second": 290.743,
      "eval_steps_per_second": 9.326,
      "eval_wer": 0.0985985985985986,
      "step": 4700
    },
    {
      "epoch": 89.62,
      "learning_rate": 7.281614543321269e-07,
      "loss": 0.7966,
      "step": 4750
    },
    {
      "epoch": 89.62,
      "eval_loss": 1.086216926574707,
      "eval_runtime": 2.2394,
      "eval_samples_per_second": 320.182,
      "eval_steps_per_second": 10.271,
      "eval_wer": 0.09617951284617951,
      "step": 4750
    },
    {
      "epoch": 90.57,
      "learning_rate": 6.098499163284066e-07,
      "loss": 0.7295,
      "step": 4800
    },
    {
      "epoch": 90.57,
      "eval_loss": 1.0895365476608276,
      "eval_runtime": 2.3225,
      "eval_samples_per_second": 308.716,
      "eval_steps_per_second": 9.903,
      "eval_wer": 0.09684684684684684,
      "step": 4800
    },
    {
      "epoch": 91.51,
      "learning_rate": 5.017204531061159e-07,
      "loss": 0.8447,
      "step": 4850
    },
    {
      "epoch": 91.51,
      "eval_loss": 1.0906962156295776,
      "eval_runtime": 2.2578,
      "eval_samples_per_second": 317.569,
      "eval_steps_per_second": 10.187,
      "eval_wer": 0.0980980980980981,
      "step": 4850
    },
    {
      "epoch": 92.45,
      "learning_rate": 4.03890313300348e-07,
      "loss": 0.7192,
      "step": 4900
    },
    {
      "epoch": 92.45,
      "eval_loss": 1.0872280597686768,
      "eval_runtime": 2.4695,
      "eval_samples_per_second": 290.337,
      "eval_steps_per_second": 9.313,
      "eval_wer": 0.09668001334668001,
      "step": 4900
    },
    {
      "epoch": 93.4,
      "learning_rate": 3.1646557762232355e-07,
      "loss": 0.8368,
      "step": 4950
    },
    {
      "epoch": 93.4,
      "eval_loss": 1.0874879360198975,
      "eval_runtime": 2.3033,
      "eval_samples_per_second": 311.286,
      "eval_steps_per_second": 9.985,
      "eval_wer": 0.0970970970970971,
      "step": 4950
    },
    {
      "epoch": 94.34,
      "learning_rate": 2.3954104383230316e-07,
      "loss": 0.7808,
      "step": 5000
    },
    {
      "epoch": 94.34,
      "eval_loss": 1.0887466669082642,
      "eval_runtime": 2.2926,
      "eval_samples_per_second": 312.748,
      "eval_steps_per_second": 10.032,
      "eval_wer": 0.09768101434768102,
      "step": 5000
    },
    {
      "epoch": 95.28,
      "learning_rate": 1.7320012394698383e-07,
      "loss": 0.76,
      "step": 5050
    },
    {
      "epoch": 95.28,
      "eval_loss": 1.089560627937317,
      "eval_runtime": 2.3151,
      "eval_samples_per_second": 309.706,
      "eval_steps_per_second": 9.935,
      "eval_wer": 0.09776443109776443,
      "step": 5050
    },
    {
      "epoch": 96.23,
      "learning_rate": 1.175147537928778e-07,
      "loss": 0.7858,
      "step": 5100
    },
    {
      "epoch": 96.23,
      "eval_loss": 1.0896259546279907,
      "eval_runtime": 2.4357,
      "eval_samples_per_second": 294.374,
      "eval_steps_per_second": 9.443,
      "eval_wer": 0.09743076409743076,
      "step": 5100
    },
    {
      "epoch": 97.17,
      "learning_rate": 7.254531500372697e-08,
      "loss": 0.766,
      "step": 5150
    },
    {
      "epoch": 97.17,
      "eval_loss": 1.0894180536270142,
      "eval_runtime": 2.3342,
      "eval_samples_per_second": 307.17,
      "eval_steps_per_second": 9.853,
      "eval_wer": 0.09776443109776443,
      "step": 5150
    },
    {
      "epoch": 98.11,
      "learning_rate": 3.8340569546546637e-08,
      "loss": 0.7899,
      "step": 5200
    },
    {
      "epoch": 98.11,
      "eval_loss": 1.0898330211639404,
      "eval_runtime": 2.3756,
      "eval_samples_per_second": 301.824,
      "eval_steps_per_second": 9.682,
      "eval_wer": 0.09776443109776443,
      "step": 5200
    },
    {
      "epoch": 99.06,
      "learning_rate": 1.4937606847278406e-08,
      "loss": 0.784,
      "step": 5250
    },
    {
      "epoch": 99.06,
      "eval_loss": 1.0888910293579102,
      "eval_runtime": 2.345,
      "eval_samples_per_second": 305.763,
      "eval_steps_per_second": 9.808,
      "eval_wer": 0.09776443109776443,
      "step": 5250
    },
    {
      "epoch": 100.0,
      "learning_rate": 2.3618035734074285e-09,
      "loss": 0.801,
      "step": 5300
    },
    {
      "epoch": 100.0,
      "eval_loss": 1.0890378952026367,
      "eval_runtime": 2.3373,
      "eval_samples_per_second": 306.769,
      "eval_steps_per_second": 9.841,
      "eval_wer": 0.09784784784784785,
      "step": 5300
    },
    {
      "epoch": 100.0,
      "step": 5300,
      "total_flos": 8.037448514378138e+16,
      "train_loss": 12.730650952177228,
      "train_runtime": 4054.5004,
      "train_samples_per_second": 83.463,
      "train_steps_per_second": 1.307
    }
  ],
  "logging_steps": 50,
  "max_steps": 5300,
  "num_train_epochs": 100,
  "save_steps": 50,
  "total_flos": 8.037448514378138e+16,
  "trial_name": null,
  "trial_params": null
}