{
  "best_metric": 49.85901151405969,
  "best_model_checkpoint": "./whisper-small-taiwanese/checkpoint-6000",
  "epoch": 2.4,
  "eval_steps": 1000,
  "global_step": 6000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "grad_norm": 202.904541015625,
      "learning_rate": 5.000000000000001e-07,
      "loss": 7.1597,
      "step": 25
    },
    {
      "epoch": 0.02,
      "grad_norm": 52.653865814208984,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 5.416,
      "step": 50
    },
    {
      "epoch": 0.03,
      "grad_norm": 32.37702178955078,
      "learning_rate": 1.5e-06,
      "loss": 3.7601,
      "step": 75
    },
    {
      "epoch": 0.04,
      "grad_norm": 29.413818359375,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 2.7107,
      "step": 100
    },
    {
      "epoch": 0.05,
      "grad_norm": 24.6357364654541,
      "learning_rate": 2.5e-06,
      "loss": 2.3941,
      "step": 125
    },
    {
      "epoch": 0.06,
      "grad_norm": 27.803171157836914,
      "learning_rate": 3e-06,
      "loss": 2.21,
      "step": 150
    },
    {
      "epoch": 0.07,
      "grad_norm": 25.927200317382812,
      "learning_rate": 3.5e-06,
      "loss": 2.0774,
      "step": 175
    },
    {
      "epoch": 0.08,
      "grad_norm": 21.307125091552734,
      "learning_rate": 4.000000000000001e-06,
      "loss": 1.8006,
      "step": 200
    },
    {
      "epoch": 0.09,
      "grad_norm": 17.714780807495117,
      "learning_rate": 4.5e-06,
      "loss": 1.598,
      "step": 225
    },
    {
      "epoch": 0.1,
      "grad_norm": 17.01283836364746,
      "learning_rate": 5e-06,
      "loss": 1.6058,
      "step": 250
    },
    {
      "epoch": 0.11,
      "grad_norm": 15.189681053161621,
      "learning_rate": 5.500000000000001e-06,
      "loss": 1.6281,
      "step": 275
    },
    {
      "epoch": 0.12,
      "grad_norm": 17.34081268310547,
      "learning_rate": 6e-06,
      "loss": 1.4216,
      "step": 300
    },
    {
      "epoch": 0.13,
      "grad_norm": 14.518415451049805,
      "learning_rate": 6.5000000000000004e-06,
      "loss": 1.5345,
      "step": 325
    },
    {
      "epoch": 0.14,
      "grad_norm": 14.931073188781738,
      "learning_rate": 7e-06,
      "loss": 1.4661,
      "step": 350
    },
    {
      "epoch": 0.15,
      "grad_norm": 15.447633743286133,
      "learning_rate": 7.500000000000001e-06,
      "loss": 1.4971,
      "step": 375
    },
    {
      "epoch": 0.16,
      "grad_norm": 15.976614952087402,
      "learning_rate": 8.000000000000001e-06,
      "loss": 1.3329,
      "step": 400
    },
    {
      "epoch": 0.17,
      "grad_norm": 14.330412864685059,
      "learning_rate": 8.5e-06,
      "loss": 1.4341,
      "step": 425
    },
    {
      "epoch": 0.18,
      "grad_norm": 14.005208969116211,
      "learning_rate": 9e-06,
      "loss": 1.3309,
      "step": 450
    },
    {
      "epoch": 0.19,
      "grad_norm": 16.984867095947266,
      "learning_rate": 9.5e-06,
      "loss": 1.4328,
      "step": 475
    },
    {
      "epoch": 0.2,
      "grad_norm": 16.296607971191406,
      "learning_rate": 1e-05,
      "loss": 1.3831,
      "step": 500
    },
    {
      "epoch": 0.21,
      "grad_norm": 14.563655853271484,
      "learning_rate": 9.944444444444445e-06,
      "loss": 1.2605,
      "step": 525
    },
    {
      "epoch": 0.22,
      "grad_norm": 15.804620742797852,
      "learning_rate": 9.88888888888889e-06,
      "loss": 1.3983,
      "step": 550
    },
    {
      "epoch": 0.23,
      "grad_norm": 14.773957252502441,
      "learning_rate": 9.833333333333333e-06,
      "loss": 1.2467,
      "step": 575
    },
    {
      "epoch": 0.24,
      "grad_norm": 14.722207069396973,
      "learning_rate": 9.777777777777779e-06,
      "loss": 1.3279,
      "step": 600
    },
    {
      "epoch": 0.25,
      "grad_norm": 14.244134902954102,
      "learning_rate": 9.722222222222223e-06,
      "loss": 1.277,
      "step": 625
    },
    {
      "epoch": 0.26,
      "grad_norm": 12.954594612121582,
      "learning_rate": 9.666666666666667e-06,
      "loss": 1.3164,
      "step": 650
    },
    {
      "epoch": 0.27,
      "grad_norm": 15.33368968963623,
      "learning_rate": 9.611111111111112e-06,
      "loss": 1.3252,
      "step": 675
    },
    {
      "epoch": 0.28,
      "grad_norm": 13.695683479309082,
      "learning_rate": 9.555555555555556e-06,
      "loss": 1.3171,
      "step": 700
    },
    {
      "epoch": 0.29,
      "grad_norm": 13.339698791503906,
      "learning_rate": 9.5e-06,
      "loss": 1.2452,
      "step": 725
    },
    {
      "epoch": 0.3,
      "grad_norm": 15.309285163879395,
      "learning_rate": 9.444444444444445e-06,
      "loss": 1.2526,
      "step": 750
    },
    {
      "epoch": 0.31,
      "grad_norm": 14.412714958190918,
      "learning_rate": 9.38888888888889e-06,
      "loss": 1.2712,
      "step": 775
    },
    {
      "epoch": 0.32,
      "grad_norm": 15.290837287902832,
      "learning_rate": 9.333333333333334e-06,
      "loss": 1.2745,
      "step": 800
    },
    {
      "epoch": 0.33,
      "grad_norm": 15.659004211425781,
      "learning_rate": 9.277777777777778e-06,
      "loss": 1.2776,
      "step": 825
    },
    {
      "epoch": 0.34,
      "grad_norm": 15.217809677124023,
      "learning_rate": 9.222222222222224e-06,
      "loss": 1.1924,
      "step": 850
    },
    {
      "epoch": 0.35,
      "grad_norm": 14.651033401489258,
      "learning_rate": 9.166666666666666e-06,
      "loss": 1.0912,
      "step": 875
    },
    {
      "epoch": 0.36,
      "grad_norm": 14.324959754943848,
      "learning_rate": 9.111111111111112e-06,
      "loss": 1.2222,
      "step": 900
    },
    {
      "epoch": 0.37,
      "grad_norm": 12.53036117553711,
      "learning_rate": 9.055555555555556e-06,
      "loss": 1.2771,
      "step": 925
    },
    {
      "epoch": 0.38,
      "grad_norm": 14.033557891845703,
      "learning_rate": 9e-06,
      "loss": 1.1506,
      "step": 950
    },
    {
      "epoch": 0.39,
      "grad_norm": 11.66818618774414,
      "learning_rate": 8.944444444444446e-06,
      "loss": 1.2423,
      "step": 975
    },
    {
      "epoch": 0.4,
      "grad_norm": 13.90597915649414,
      "learning_rate": 8.888888888888888e-06,
      "loss": 1.1421,
      "step": 1000
    },
    {
      "epoch": 0.4,
      "eval_cer": 61.163904814262146,
      "eval_loss": 1.1691664457321167,
      "eval_runtime": 1744.2126,
      "eval_samples_per_second": 2.257,
      "eval_steps_per_second": 0.282,
      "step": 1000
    },
    {
      "epoch": 0.41,
      "grad_norm": 11.293231964111328,
      "learning_rate": 8.833333333333334e-06,
      "loss": 1.153,
      "step": 1025
    },
    {
      "epoch": 0.42,
      "grad_norm": 13.314165115356445,
      "learning_rate": 8.777777777777778e-06,
      "loss": 1.168,
      "step": 1050
    },
    {
      "epoch": 0.43,
      "grad_norm": 13.231385231018066,
      "learning_rate": 8.722222222222224e-06,
      "loss": 1.1613,
      "step": 1075
    },
    {
      "epoch": 0.44,
      "grad_norm": 13.21717643737793,
      "learning_rate": 8.666666666666668e-06,
      "loss": 1.1246,
      "step": 1100
    },
    {
      "epoch": 0.45,
      "grad_norm": 11.046935081481934,
      "learning_rate": 8.611111111111112e-06,
      "loss": 1.088,
      "step": 1125
    },
    {
      "epoch": 0.46,
      "grad_norm": 14.906622886657715,
      "learning_rate": 8.555555555555556e-06,
      "loss": 1.19,
      "step": 1150
    },
    {
      "epoch": 0.47,
      "grad_norm": 14.302517890930176,
      "learning_rate": 8.5e-06,
      "loss": 1.1351,
      "step": 1175
    },
    {
      "epoch": 0.48,
      "grad_norm": 13.947770118713379,
      "learning_rate": 8.444444444444446e-06,
      "loss": 1.057,
      "step": 1200
    },
    {
      "epoch": 0.49,
      "grad_norm": 14.45609188079834,
      "learning_rate": 8.38888888888889e-06,
      "loss": 1.0993,
      "step": 1225
    },
    {
      "epoch": 0.5,
      "grad_norm": 14.952827453613281,
      "learning_rate": 8.333333333333334e-06,
      "loss": 1.1626,
      "step": 1250
    },
    {
      "epoch": 0.51,
      "grad_norm": 16.128353118896484,
      "learning_rate": 8.277777777777778e-06,
      "loss": 1.1082,
      "step": 1275
    },
    {
      "epoch": 0.52,
      "grad_norm": 13.550396919250488,
      "learning_rate": 8.222222222222222e-06,
      "loss": 1.183,
      "step": 1300
    },
    {
      "epoch": 0.53,
      "grad_norm": 14.400228500366211,
      "learning_rate": 8.166666666666668e-06,
      "loss": 1.1988,
      "step": 1325
    },
    {
      "epoch": 0.54,
      "grad_norm": 13.9801607131958,
      "learning_rate": 8.111111111111112e-06,
      "loss": 1.1314,
      "step": 1350
    },
    {
      "epoch": 0.55,
      "grad_norm": 12.84874439239502,
      "learning_rate": 8.055555555555557e-06,
      "loss": 1.1411,
      "step": 1375
    },
    {
      "epoch": 0.56,
      "grad_norm": 14.126324653625488,
      "learning_rate": 8.000000000000001e-06,
      "loss": 1.1314,
      "step": 1400
    },
    {
      "epoch": 0.57,
      "grad_norm": 12.402750015258789,
      "learning_rate": 7.944444444444445e-06,
      "loss": 1.1071,
      "step": 1425
    },
    {
      "epoch": 0.58,
      "grad_norm": 13.835284233093262,
      "learning_rate": 7.88888888888889e-06,
      "loss": 1.1393,
      "step": 1450
    },
    {
      "epoch": 0.59,
      "grad_norm": 12.414569854736328,
      "learning_rate": 7.833333333333333e-06,
      "loss": 1.1026,
      "step": 1475
    },
    {
      "epoch": 0.6,
      "grad_norm": 15.43626880645752,
      "learning_rate": 7.77777777777778e-06,
      "loss": 1.164,
      "step": 1500
    },
    {
      "epoch": 0.61,
      "grad_norm": 13.067487716674805,
      "learning_rate": 7.722222222222223e-06,
      "loss": 1.0448,
      "step": 1525
    },
    {
      "epoch": 0.62,
      "grad_norm": 14.158551216125488,
      "learning_rate": 7.666666666666667e-06,
      "loss": 1.1674,
      "step": 1550
    },
    {
      "epoch": 0.63,
      "grad_norm": 13.062005996704102,
      "learning_rate": 7.611111111111111e-06,
      "loss": 1.0916,
      "step": 1575
    },
    {
      "epoch": 0.64,
      "grad_norm": 13.6104736328125,
      "learning_rate": 7.555555555555556e-06,
      "loss": 1.0424,
      "step": 1600
    },
    {
      "epoch": 0.65,
      "grad_norm": 11.52835750579834,
      "learning_rate": 7.500000000000001e-06,
      "loss": 1.0196,
      "step": 1625
    },
    {
      "epoch": 0.66,
      "grad_norm": 14.118935585021973,
      "learning_rate": 7.444444444444445e-06,
      "loss": 1.1502,
      "step": 1650
    },
    {
      "epoch": 0.67,
      "grad_norm": 13.2473726272583,
      "learning_rate": 7.38888888888889e-06,
      "loss": 1.0562,
      "step": 1675
    },
    {
      "epoch": 0.68,
      "grad_norm": 13.026944160461426,
      "learning_rate": 7.333333333333333e-06,
      "loss": 1.0391,
      "step": 1700
    },
    {
      "epoch": 0.69,
      "grad_norm": 11.923539161682129,
      "learning_rate": 7.277777777777778e-06,
      "loss": 1.063,
      "step": 1725
    },
    {
      "epoch": 0.7,
      "grad_norm": 10.581581115722656,
      "learning_rate": 7.222222222222223e-06,
      "loss": 1.0504,
      "step": 1750
    },
    {
      "epoch": 0.71,
      "grad_norm": 13.761798858642578,
      "learning_rate": 7.166666666666667e-06,
      "loss": 1.1781,
      "step": 1775
    },
    {
      "epoch": 0.72,
      "grad_norm": 13.440286636352539,
      "learning_rate": 7.111111111111112e-06,
      "loss": 1.088,
      "step": 1800
    },
    {
      "epoch": 0.73,
      "grad_norm": 11.378331184387207,
      "learning_rate": 7.055555555555557e-06,
      "loss": 1.017,
      "step": 1825
    },
    {
      "epoch": 0.74,
      "grad_norm": 16.24916648864746,
      "learning_rate": 7e-06,
      "loss": 1.0669,
      "step": 1850
    },
    {
      "epoch": 0.75,
      "grad_norm": 14.499041557312012,
      "learning_rate": 6.944444444444445e-06,
      "loss": 1.023,
      "step": 1875
    },
    {
      "epoch": 0.76,
      "grad_norm": 14.587787628173828,
      "learning_rate": 6.88888888888889e-06,
      "loss": 1.1128,
      "step": 1900
    },
    {
      "epoch": 0.77,
      "grad_norm": 14.249890327453613,
      "learning_rate": 6.833333333333334e-06,
      "loss": 1.0462,
      "step": 1925
    },
    {
      "epoch": 0.78,
      "grad_norm": 13.22544002532959,
      "learning_rate": 6.777777777777779e-06,
      "loss": 1.0564,
      "step": 1950
    },
    {
      "epoch": 0.79,
      "grad_norm": 13.404162406921387,
      "learning_rate": 6.7222222222222235e-06,
      "loss": 1.0517,
      "step": 1975
    },
    {
      "epoch": 0.8,
      "grad_norm": 13.87370491027832,
      "learning_rate": 6.666666666666667e-06,
      "loss": 1.0556,
      "step": 2000
    },
    {
      "epoch": 0.8,
      "eval_cer": 51.77491557370612,
      "eval_loss": 1.0214924812316895,
      "eval_runtime": 1738.5549,
      "eval_samples_per_second": 2.264,
      "eval_steps_per_second": 0.283,
      "step": 2000
    },
    {
      "epoch": 0.81,
      "grad_norm": 15.01921272277832,
      "learning_rate": 6.6111111111111115e-06,
      "loss": 1.0203,
      "step": 2025
    },
    {
      "epoch": 0.82,
      "grad_norm": 12.810806274414062,
      "learning_rate": 6.555555555555556e-06,
      "loss": 1.0426,
      "step": 2050
    },
    {
      "epoch": 0.83,
      "grad_norm": 12.653311729431152,
      "learning_rate": 6.5000000000000004e-06,
      "loss": 1.0451,
      "step": 2075
    },
    {
      "epoch": 0.84,
      "grad_norm": 14.353548049926758,
      "learning_rate": 6.444444444444445e-06,
      "loss": 1.0853,
      "step": 2100
    },
    {
      "epoch": 0.85,
      "grad_norm": 12.075662612915039,
      "learning_rate": 6.3888888888888885e-06,
      "loss": 1.0124,
      "step": 2125
    },
    {
      "epoch": 0.86,
      "grad_norm": 14.976198196411133,
      "learning_rate": 6.333333333333333e-06,
      "loss": 1.0461,
      "step": 2150
    },
    {
      "epoch": 0.87,
      "grad_norm": 12.702341079711914,
      "learning_rate": 6.277777777777778e-06,
      "loss": 1.0319,
      "step": 2175
    },
    {
      "epoch": 0.88,
      "grad_norm": 14.232624053955078,
      "learning_rate": 6.222222222222223e-06,
      "loss": 1.0174,
      "step": 2200
    },
    {
      "epoch": 0.89,
      "grad_norm": 12.84585189819336,
      "learning_rate": 6.166666666666667e-06,
      "loss": 1.0275,
      "step": 2225
    },
    {
      "epoch": 0.9,
      "grad_norm": 17.438783645629883,
      "learning_rate": 6.111111111111112e-06,
      "loss": 0.9808,
      "step": 2250
    },
    {
      "epoch": 0.91,
      "grad_norm": 13.450716018676758,
      "learning_rate": 6.055555555555555e-06,
      "loss": 1.0239,
      "step": 2275
    },
    {
      "epoch": 0.92,
      "grad_norm": 14.077709197998047,
      "learning_rate": 6e-06,
      "loss": 1.0151,
      "step": 2300
    },
    {
      "epoch": 0.93,
      "grad_norm": 10.755902290344238,
      "learning_rate": 5.944444444444445e-06,
      "loss": 0.9994,
      "step": 2325
    },
    {
      "epoch": 0.94,
      "grad_norm": 13.025961875915527,
      "learning_rate": 5.88888888888889e-06,
      "loss": 1.0136,
      "step": 2350
    },
    {
      "epoch": 0.95,
      "grad_norm": 13.124420166015625,
      "learning_rate": 5.833333333333334e-06,
      "loss": 0.9732,
      "step": 2375
    },
    {
      "epoch": 0.96,
      "grad_norm": 13.470574378967285,
      "learning_rate": 5.777777777777778e-06,
      "loss": 0.9919,
      "step": 2400
    },
    {
      "epoch": 0.97,
      "grad_norm": 13.073882102966309,
      "learning_rate": 5.722222222222222e-06,
      "loss": 0.9791,
      "step": 2425
    },
    {
      "epoch": 0.98,
      "grad_norm": 12.156904220581055,
      "learning_rate": 5.666666666666667e-06,
      "loss": 0.9761,
      "step": 2450
    },
    {
      "epoch": 0.99,
      "grad_norm": 14.560174942016602,
      "learning_rate": 5.611111111111112e-06,
      "loss": 0.9478,
      "step": 2475
    },
    {
      "epoch": 1.0,
      "grad_norm": 13.365988731384277,
      "learning_rate": 5.555555555555557e-06,
      "loss": 0.9439,
      "step": 2500
    },
    {
      "epoch": 1.01,
      "grad_norm": 11.116703987121582,
      "learning_rate": 5.500000000000001e-06,
      "loss": 0.8144,
      "step": 2525
    },
    {
      "epoch": 1.02,
      "grad_norm": 11.22291088104248,
      "learning_rate": 5.444444444444445e-06,
      "loss": 0.7626,
      "step": 2550
    },
    {
      "epoch": 1.03,
      "grad_norm": 10.895951271057129,
      "learning_rate": 5.388888888888889e-06,
      "loss": 0.7443,
      "step": 2575
    },
    {
      "epoch": 1.04,
      "grad_norm": 10.497587203979492,
      "learning_rate": 5.333333333333334e-06,
      "loss": 0.7348,
      "step": 2600
    },
    {
      "epoch": 1.05,
      "grad_norm": 10.018071174621582,
      "learning_rate": 5.2777777777777785e-06,
      "loss": 0.7096,
      "step": 2625
    },
    {
      "epoch": 1.06,
      "grad_norm": 11.595548629760742,
      "learning_rate": 5.2222222222222226e-06,
      "loss": 0.7527,
      "step": 2650
    },
    {
      "epoch": 1.07,
      "grad_norm": 9.90001106262207,
      "learning_rate": 5.1666666666666675e-06,
      "loss": 0.7502,
      "step": 2675
    },
    {
      "epoch": 1.08,
      "grad_norm": 14.49976921081543,
      "learning_rate": 5.1111111111111115e-06,
      "loss": 0.7253,
      "step": 2700
    },
    {
      "epoch": 1.09,
      "grad_norm": 10.992387771606445,
      "learning_rate": 5.0555555555555555e-06,
      "loss": 0.759,
      "step": 2725
    },
    {
      "epoch": 1.1,
      "grad_norm": 11.267956733703613,
      "learning_rate": 5e-06,
      "loss": 0.7174,
      "step": 2750
    },
    {
      "epoch": 1.11,
      "grad_norm": 12.558463096618652,
      "learning_rate": 4.944444444444445e-06,
      "loss": 0.7936,
      "step": 2775
    },
    {
      "epoch": 1.12,
      "grad_norm": 10.136383056640625,
      "learning_rate": 4.888888888888889e-06,
      "loss": 0.7573,
      "step": 2800
    },
    {
      "epoch": 1.13,
      "grad_norm": 10.183345794677734,
      "learning_rate": 4.833333333333333e-06,
      "loss": 0.7704,
      "step": 2825
    },
    {
      "epoch": 1.1400000000000001,
      "grad_norm": 10.488049507141113,
      "learning_rate": 4.777777777777778e-06,
      "loss": 0.7343,
      "step": 2850
    },
    {
      "epoch": 1.15,
      "grad_norm": 12.607866287231445,
      "learning_rate": 4.722222222222222e-06,
      "loss": 0.7247,
      "step": 2875
    },
    {
      "epoch": 1.16,
      "grad_norm": 10.342514991760254,
      "learning_rate": 4.666666666666667e-06,
      "loss": 0.7618,
      "step": 2900
    },
    {
      "epoch": 1.17,
      "grad_norm": 10.273557662963867,
      "learning_rate": 4.611111111111112e-06,
      "loss": 0.805,
      "step": 2925
    },
    {
      "epoch": 1.18,
      "grad_norm": 11.328307151794434,
      "learning_rate": 4.555555555555556e-06,
      "loss": 0.6998,
      "step": 2950
    },
    {
      "epoch": 1.19,
      "grad_norm": 13.699926376342773,
      "learning_rate": 4.5e-06,
      "loss": 0.7924,
      "step": 2975
    },
    {
      "epoch": 1.2,
      "grad_norm": 10.797248840332031,
      "learning_rate": 4.444444444444444e-06,
      "loss": 0.7837,
      "step": 3000
    },
    {
      "epoch": 1.2,
      "eval_cer": 54.1977538679023,
      "eval_loss": 0.9572013020515442,
      "eval_runtime": 1756.4904,
      "eval_samples_per_second": 2.241,
      "eval_steps_per_second": 0.28,
      "step": 3000
    },
    {
      "epoch": 1.21,
      "grad_norm": 12.410914421081543,
      "learning_rate": 4.388888888888889e-06,
      "loss": 0.7325,
      "step": 3025
    },
    {
      "epoch": 1.22,
      "grad_norm": 12.31027603149414,
      "learning_rate": 4.333333333333334e-06,
      "loss": 0.6914,
      "step": 3050
    },
    {
      "epoch": 1.23,
      "grad_norm": 10.667587280273438,
      "learning_rate": 4.277777777777778e-06,
      "loss": 0.7216,
      "step": 3075
    },
    {
      "epoch": 1.24,
      "grad_norm": 13.815753936767578,
      "learning_rate": 4.222222222222223e-06,
      "loss": 0.7235,
      "step": 3100
    },
    {
      "epoch": 1.25,
      "grad_norm": 11.412848472595215,
      "learning_rate": 4.166666666666667e-06,
      "loss": 0.7579,
      "step": 3125
    },
    {
      "epoch": 1.26,
      "grad_norm": 13.73943042755127,
      "learning_rate": 4.111111111111111e-06,
      "loss": 0.77,
      "step": 3150
    },
    {
      "epoch": 1.27,
      "grad_norm": 13.075247764587402,
      "learning_rate": 4.055555555555556e-06,
      "loss": 0.6957,
      "step": 3175
    },
    {
      "epoch": 1.28,
      "grad_norm": 11.140027046203613,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.7516,
      "step": 3200
    },
    {
      "epoch": 1.29,
      "grad_norm": 13.179604530334473,
      "learning_rate": 3.944444444444445e-06,
      "loss": 0.7427,
      "step": 3225
    },
    {
      "epoch": 1.3,
      "grad_norm": 11.4240140914917,
      "learning_rate": 3.88888888888889e-06,
      "loss": 0.7679,
      "step": 3250
    },
    {
      "epoch": 1.31,
      "grad_norm": 10.857213020324707,
      "learning_rate": 3.833333333333334e-06,
      "loss": 0.69,
      "step": 3275
    },
    {
      "epoch": 1.32,
      "grad_norm": 12.246594429016113,
      "learning_rate": 3.777777777777778e-06,
      "loss": 0.7841,
      "step": 3300
    },
    {
      "epoch": 1.33,
      "grad_norm": 10.60627269744873,
      "learning_rate": 3.7222222222222225e-06,
      "loss": 0.7167,
      "step": 3325
    },
    {
      "epoch": 1.34,
      "grad_norm": 15.023033142089844,
      "learning_rate": 3.6666666666666666e-06,
      "loss": 0.753,
      "step": 3350
    },
    {
      "epoch": 1.35,
      "grad_norm": 11.280016899108887,
      "learning_rate": 3.6111111111111115e-06,
      "loss": 0.7304,
      "step": 3375
    },
    {
      "epoch": 1.3599999999999999,
      "grad_norm": 12.962152481079102,
      "learning_rate": 3.555555555555556e-06,
      "loss": 0.7802,
      "step": 3400
    },
    {
      "epoch": 1.37,
      "grad_norm": 11.989816665649414,
      "learning_rate": 3.5e-06,
      "loss": 0.7298,
      "step": 3425
    },
    {
      "epoch": 1.38,
      "grad_norm": 11.917997360229492,
      "learning_rate": 3.444444444444445e-06,
      "loss": 0.8122,
      "step": 3450
    },
    {
      "epoch": 1.3900000000000001,
      "grad_norm": 11.995168685913086,
      "learning_rate": 3.3888888888888893e-06,
      "loss": 0.7428,
      "step": 3475
    },
    {
      "epoch": 1.4,
      "grad_norm": 13.24657917022705,
      "learning_rate": 3.3333333333333333e-06,
      "loss": 0.7487,
      "step": 3500
    },
    {
      "epoch": 1.41,
      "grad_norm": 14.055580139160156,
      "learning_rate": 3.277777777777778e-06,
      "loss": 0.735,
      "step": 3525
    },
    {
      "epoch": 1.42,
      "grad_norm": 11.190061569213867,
      "learning_rate": 3.2222222222222227e-06,
      "loss": 0.747,
      "step": 3550
    },
    {
      "epoch": 1.43,
      "grad_norm": 10.754741668701172,
      "learning_rate": 3.1666666666666667e-06,
      "loss": 0.7953,
      "step": 3575
    },
    {
      "epoch": 1.44,
      "grad_norm": 11.28659725189209,
      "learning_rate": 3.1111111111111116e-06,
      "loss": 0.7396,
      "step": 3600
    },
    {
      "epoch": 1.45,
      "grad_norm": 9.70627498626709,
      "learning_rate": 3.055555555555556e-06,
      "loss": 0.7127,
      "step": 3625
    },
    {
      "epoch": 1.46,
      "grad_norm": 13.152200698852539,
      "learning_rate": 3e-06,
      "loss": 0.7299,
      "step": 3650
    },
    {
      "epoch": 1.47,
      "grad_norm": 12.44729995727539,
      "learning_rate": 2.944444444444445e-06,
      "loss": 0.7195,
      "step": 3675
    },
    {
      "epoch": 1.48,
      "grad_norm": 10.643758773803711,
      "learning_rate": 2.888888888888889e-06,
      "loss": 0.7772,
      "step": 3700
    },
    {
      "epoch": 1.49,
      "grad_norm": 13.989819526672363,
      "learning_rate": 2.8333333333333335e-06,
      "loss": 0.7405,
      "step": 3725
    },
    {
      "epoch": 1.5,
      "grad_norm": 9.972952842712402,
      "learning_rate": 2.7777777777777783e-06,
      "loss": 0.722,
      "step": 3750
    },
    {
      "epoch": 1.51,
      "grad_norm": 12.274140357971191,
      "learning_rate": 2.7222222222222224e-06,
      "loss": 0.7326,
      "step": 3775
    },
    {
      "epoch": 1.52,
      "grad_norm": 10.712841033935547,
      "learning_rate": 2.666666666666667e-06,
      "loss": 0.6631,
      "step": 3800
    },
    {
      "epoch": 1.53,
      "grad_norm": 13.03647232055664,
      "learning_rate": 2.6111111111111113e-06,
      "loss": 0.6943,
      "step": 3825
    },
    {
      "epoch": 1.54,
      "grad_norm": 10.963674545288086,
      "learning_rate": 2.5555555555555557e-06,
      "loss": 0.7379,
      "step": 3850
    },
    {
      "epoch": 1.55,
      "grad_norm": 11.267337799072266,
      "learning_rate": 2.5e-06,
      "loss": 0.7165,
      "step": 3875
    },
    {
      "epoch": 1.56,
      "grad_norm": 10.950325965881348,
      "learning_rate": 2.4444444444444447e-06,
      "loss": 0.718,
      "step": 3900
    },
    {
      "epoch": 1.5699999999999998,
      "grad_norm": 12.390423774719238,
      "learning_rate": 2.388888888888889e-06,
      "loss": 0.76,
      "step": 3925
    },
    {
      "epoch": 1.58,
      "grad_norm": 11.487335205078125,
      "learning_rate": 2.3333333333333336e-06,
      "loss": 0.7708,
      "step": 3950
    },
    {
      "epoch": 1.5899999999999999,
      "grad_norm": 11.19138240814209,
      "learning_rate": 2.277777777777778e-06,
      "loss": 0.6526,
      "step": 3975
    },
    {
      "epoch": 1.6,
      "grad_norm": 10.622859954833984,
      "learning_rate": 2.222222222222222e-06,
      "loss": 0.7332,
      "step": 4000
    },
    {
      "epoch": 1.6,
      "eval_cer": 50.39660724102725,
      "eval_loss": 0.9230387210845947,
      "eval_runtime": 1742.867,
      "eval_samples_per_second": 2.258,
      "eval_steps_per_second": 0.282,
      "step": 4000
    },
    {
      "epoch": 1.6099999999999999,
      "grad_norm": 12.709884643554688,
      "learning_rate": 2.166666666666667e-06,
      "loss": 0.7753,
      "step": 4025
    },
    {
      "epoch": 1.62,
      "grad_norm": 11.723743438720703,
      "learning_rate": 2.1111111111111114e-06,
      "loss": 0.6813,
      "step": 4050
    },
    {
      "epoch": 1.63,
      "grad_norm": 13.305733680725098,
      "learning_rate": 2.0555555555555555e-06,
      "loss": 0.8031,
      "step": 4075
    },
    {
      "epoch": 1.6400000000000001,
      "grad_norm": 11.093185424804688,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.6939,
      "step": 4100
    },
    {
      "epoch": 1.65,
      "grad_norm": 11.894206047058105,
      "learning_rate": 1.944444444444445e-06,
      "loss": 0.7193,
      "step": 4125
    },
    {
      "epoch": 1.6600000000000001,
      "grad_norm": 14.037126541137695,
      "learning_rate": 1.888888888888889e-06,
      "loss": 0.7113,
      "step": 4150
    },
    {
      "epoch": 1.67,
      "grad_norm": 8.205574035644531,
      "learning_rate": 1.8333333333333333e-06,
      "loss": 0.7197,
      "step": 4175
    },
    {
      "epoch": 1.6800000000000002,
      "grad_norm": 13.138123512268066,
      "learning_rate": 1.777777777777778e-06,
      "loss": 0.7747,
      "step": 4200
    },
    {
      "epoch": 1.69,
      "grad_norm": 12.72310733795166,
      "learning_rate": 1.7222222222222224e-06,
      "loss": 0.7181,
      "step": 4225
    },
    {
      "epoch": 1.7,
      "grad_norm": 12.127406120300293,
      "learning_rate": 1.6666666666666667e-06,
      "loss": 0.6518,
      "step": 4250
    },
    {
      "epoch": 1.71,
      "grad_norm": 12.796150207519531,
      "learning_rate": 1.6111111111111113e-06,
      "loss": 0.6808,
      "step": 4275
    },
    {
      "epoch": 1.72,
      "grad_norm": 9.223761558532715,
      "learning_rate": 1.5555555555555558e-06,
      "loss": 0.691,
      "step": 4300
    },
    {
      "epoch": 1.73,
      "grad_norm": 11.86072826385498,
      "learning_rate": 1.5e-06,
      "loss": 0.7387,
      "step": 4325
    },
    {
      "epoch": 1.74,
      "grad_norm": 13.236778259277344,
      "learning_rate": 1.4444444444444445e-06,
      "loss": 0.7341,
      "step": 4350
    },
    {
      "epoch": 1.75,
      "grad_norm": 12.484067916870117,
      "learning_rate": 1.3888888888888892e-06,
      "loss": 0.6508,
      "step": 4375
    },
    {
      "epoch": 1.76,
      "grad_norm": 9.531938552856445,
      "learning_rate": 1.3333333333333334e-06,
      "loss": 0.7072,
      "step": 4400
    },
    {
      "epoch": 1.77,
      "grad_norm": 12.968701362609863,
      "learning_rate": 1.2777777777777779e-06,
      "loss": 0.7091,
      "step": 4425
    },
    {
      "epoch": 1.78,
      "grad_norm": 11.115565299987793,
      "learning_rate": 1.2222222222222223e-06,
      "loss": 0.7082,
      "step": 4450
    },
    {
      "epoch": 1.79,
      "grad_norm": 15.03991985321045,
      "learning_rate": 1.1666666666666668e-06,
      "loss": 0.6707,
      "step": 4475
    },
    {
      "epoch": 1.8,
      "grad_norm": 13.241722106933594,
      "learning_rate": 1.111111111111111e-06,
      "loss": 0.7449,
      "step": 4500
    },
    {
      "epoch": 1.81,
      "grad_norm": 12.832565307617188,
      "learning_rate": 1.0555555555555557e-06,
      "loss": 0.7065,
      "step": 4525
    },
    {
      "epoch": 1.8199999999999998,
      "grad_norm": 13.823897361755371,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.7243,
      "step": 4550
    },
    {
      "epoch": 1.83,
      "grad_norm": 8.99041748046875,
      "learning_rate": 9.444444444444445e-07,
      "loss": 0.7175,
      "step": 4575
    },
    {
      "epoch": 1.8399999999999999,
      "grad_norm": 10.291349411010742,
      "learning_rate": 8.88888888888889e-07,
      "loss": 0.6431,
      "step": 4600
    },
    {
      "epoch": 1.85,
      "grad_norm": 10.827986717224121,
      "learning_rate": 8.333333333333333e-07,
      "loss": 0.6776,
      "step": 4625
    },
    {
      "epoch": 1.8599999999999999,
      "grad_norm": 11.811609268188477,
      "learning_rate": 7.777777777777779e-07,
      "loss": 0.7021,
      "step": 4650
    },
    {
      "epoch": 1.87,
      "grad_norm": 10.870746612548828,
      "learning_rate": 7.222222222222222e-07,
      "loss": 0.7036,
      "step": 4675
    },
    {
      "epoch": 1.88,
      "grad_norm": 12.761141777038574,
      "learning_rate": 6.666666666666667e-07,
      "loss": 0.6818,
      "step": 4700
    },
    {
      "epoch": 1.8900000000000001,
      "grad_norm": 12.68936538696289,
      "learning_rate": 6.111111111111112e-07,
      "loss": 0.7104,
      "step": 4725
    },
    {
      "epoch": 1.9,
      "grad_norm": 12.023513793945312,
      "learning_rate": 5.555555555555555e-07,
      "loss": 0.6955,
      "step": 4750
    },
    {
      "epoch": 1.9100000000000001,
      "grad_norm": 13.548583984375,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.7529,
      "step": 4775
    },
    {
      "epoch": 1.92,
      "grad_norm": 11.201909065246582,
      "learning_rate": 4.444444444444445e-07,
      "loss": 0.6256,
      "step": 4800
    },
    {
      "epoch": 1.9300000000000002,
      "grad_norm": 12.877849578857422,
      "learning_rate": 3.8888888888888895e-07,
      "loss": 0.7399,
      "step": 4825
    },
    {
      "epoch": 1.94,
      "grad_norm": 12.711626052856445,
      "learning_rate": 3.3333333333333335e-07,
      "loss": 0.6784,
      "step": 4850
    },
    {
      "epoch": 1.95,
      "grad_norm": 9.75576114654541,
      "learning_rate": 2.7777777777777776e-07,
      "loss": 0.7111,
      "step": 4875
    },
    {
      "epoch": 1.96,
      "grad_norm": 11.684128761291504,
      "learning_rate": 2.2222222222222224e-07,
      "loss": 0.7221,
      "step": 4900
    },
    {
      "epoch": 1.97,
      "grad_norm": 12.131952285766602,
      "learning_rate": 1.6666666666666668e-07,
      "loss": 0.7624,
      "step": 4925
    },
    {
      "epoch": 1.98,
      "grad_norm": 10.875662803649902,
      "learning_rate": 1.1111111111111112e-07,
      "loss": 0.6913,
      "step": 4950
    },
    {
      "epoch": 1.99,
      "grad_norm": 10.225822448730469,
      "learning_rate": 5.555555555555556e-08,
      "loss": 0.6974,
      "step": 4975
    },
    {
      "epoch": 2.0,
      "grad_norm": 9.928595542907715,
      "learning_rate": 0.0,
      "loss": 0.6957,
      "step": 5000
    },
    {
      "epoch": 2.0,
      "eval_cer": 50.577240241891154,
      "eval_loss": 0.9063901305198669,
      "eval_runtime": 1752.6784,
      "eval_samples_per_second": 2.246,
      "eval_steps_per_second": 0.281,
      "step": 5000
    },
    {
      "epoch": 2.01,
      "grad_norm": 12.426376342773438,
      "learning_rate": 4.907407407407408e-06,
      "loss": 0.5823,
      "step": 5025
    },
    {
      "epoch": 2.02,
      "grad_norm": 10.951026916503906,
      "learning_rate": 4.876543209876544e-06,
      "loss": 0.6085,
      "step": 5050
    },
    {
      "epoch": 2.03,
      "grad_norm": 10.53141975402832,
      "learning_rate": 4.845679012345679e-06,
      "loss": 0.5771,
      "step": 5075
    },
    {
      "epoch": 2.04,
      "grad_norm": 13.22917366027832,
      "learning_rate": 4.814814814814815e-06,
      "loss": 0.6013,
      "step": 5100
    },
    {
      "epoch": 2.05,
      "grad_norm": 12.245888710021973,
      "learning_rate": 4.783950617283951e-06,
      "loss": 0.559,
      "step": 5125
    },
    {
      "epoch": 2.06,
      "grad_norm": 10.822369575500488,
      "learning_rate": 4.753086419753087e-06,
      "loss": 0.6044,
      "step": 5150
    },
    {
      "epoch": 2.07,
      "grad_norm": 13.181960105895996,
      "learning_rate": 4.722222222222222e-06,
      "loss": 0.6063,
      "step": 5175
    },
    {
      "epoch": 2.08,
      "grad_norm": 10.481088638305664,
      "learning_rate": 4.691358024691358e-06,
      "loss": 0.572,
      "step": 5200
    },
    {
      "epoch": 2.09,
      "grad_norm": 10.6892671585083,
      "learning_rate": 4.660493827160494e-06,
      "loss": 0.6041,
      "step": 5225
    },
    {
      "epoch": 2.1,
      "grad_norm": 11.274572372436523,
      "learning_rate": 4.62962962962963e-06,
      "loss": 0.5608,
      "step": 5250
    },
    {
      "epoch": 2.11,
      "grad_norm": 11.774944305419922,
      "learning_rate": 4.598765432098766e-06,
      "loss": 0.6163,
      "step": 5275
    },
    {
      "epoch": 2.12,
      "grad_norm": 10.94110107421875,
      "learning_rate": 4.567901234567902e-06,
      "loss": 0.5816,
      "step": 5300
    },
    {
      "epoch": 2.13,
      "grad_norm": 10.177285194396973,
      "learning_rate": 4.537037037037038e-06,
      "loss": 0.557,
      "step": 5325
    },
    {
      "epoch": 2.14,
      "grad_norm": 9.426462173461914,
      "learning_rate": 4.506172839506173e-06,
      "loss": 0.6411,
      "step": 5350
    },
    {
      "epoch": 2.15,
      "grad_norm": 9.961886405944824,
      "learning_rate": 4.475308641975309e-06,
      "loss": 0.5906,
      "step": 5375
    },
    {
      "epoch": 2.16,
      "grad_norm": 11.078606605529785,
      "learning_rate": 4.444444444444444e-06,
      "loss": 0.622,
      "step": 5400
    },
    {
      "epoch": 2.17,
      "grad_norm": 10.944733619689941,
      "learning_rate": 4.413580246913581e-06,
      "loss": 0.6154,
      "step": 5425
    },
    {
      "epoch": 2.18,
      "grad_norm": 10.523876190185547,
      "learning_rate": 4.382716049382716e-06,
      "loss": 0.6698,
      "step": 5450
    },
    {
      "epoch": 2.19,
      "grad_norm": 12.090290069580078,
      "learning_rate": 4.351851851851852e-06,
      "loss": 0.6269,
      "step": 5475
    },
    {
      "epoch": 2.2,
      "grad_norm": 9.12374210357666,
      "learning_rate": 4.3209876543209875e-06,
      "loss": 0.6209,
      "step": 5500
    },
    {
      "epoch": 2.21,
      "grad_norm": 11.668550491333008,
      "learning_rate": 4.290123456790124e-06,
      "loss": 0.6692,
      "step": 5525
    },
    {
      "epoch": 2.22,
      "grad_norm": 13.481975555419922,
      "learning_rate": 4.2592592592592596e-06,
      "loss": 0.6093,
      "step": 5550
    },
    {
      "epoch": 2.23,
      "grad_norm": 12.388972282409668,
      "learning_rate": 4.228395061728396e-06,
      "loss": 0.5941,
      "step": 5575
    },
    {
      "epoch": 2.24,
      "grad_norm": 9.559195518493652,
      "learning_rate": 4.197530864197531e-06,
      "loss": 0.5782,
      "step": 5600
    },
    {
      "epoch": 2.25,
      "grad_norm": 10.876689910888672,
      "learning_rate": 4.166666666666667e-06,
      "loss": 0.5896,
      "step": 5625
    },
    {
      "epoch": 2.26,
      "grad_norm": 10.03794002532959,
      "learning_rate": 4.135802469135803e-06,
      "loss": 0.6243,
      "step": 5650
    },
    {
      "epoch": 2.27,
      "grad_norm": 11.717126846313477,
      "learning_rate": 4.104938271604938e-06,
      "loss": 0.6205,
      "step": 5675
    },
    {
      "epoch": 2.2800000000000002,
      "grad_norm": 13.434377670288086,
      "learning_rate": 4.074074074074074e-06,
      "loss": 0.6419,
      "step": 5700
    },
    {
      "epoch": 2.29,
      "grad_norm": 8.062933921813965,
      "learning_rate": 4.04320987654321e-06,
      "loss": 0.6042,
      "step": 5725
    },
    {
      "epoch": 2.3,
      "grad_norm": 11.512144088745117,
      "learning_rate": 4.012345679012346e-06,
      "loss": 0.5766,
      "step": 5750
    },
    {
      "epoch": 2.31,
      "grad_norm": 12.437589645385742,
      "learning_rate": 3.9814814814814814e-06,
      "loss": 0.6139,
      "step": 5775
    },
    {
      "epoch": 2.32,
      "grad_norm": 13.863161087036133,
      "learning_rate": 3.9506172839506175e-06,
      "loss": 0.6477,
      "step": 5800
    },
    {
      "epoch": 2.33,
      "grad_norm": 12.266985893249512,
      "learning_rate": 3.9197530864197535e-06,
      "loss": 0.5743,
      "step": 5825
    },
    {
      "epoch": 2.34,
      "grad_norm": 9.438658714294434,
      "learning_rate": 3.88888888888889e-06,
      "loss": 0.5324,
      "step": 5850
    },
    {
      "epoch": 2.35,
      "grad_norm": 10.063249588012695,
      "learning_rate": 3.858024691358025e-06,
      "loss": 0.5516,
      "step": 5875
    },
    {
      "epoch": 2.36,
      "grad_norm": 11.568090438842773,
      "learning_rate": 3.827160493827161e-06,
      "loss": 0.6009,
      "step": 5900
    },
    {
      "epoch": 2.37,
      "grad_norm": 12.200023651123047,
      "learning_rate": 3.796296296296297e-06,
      "loss": 0.5942,
      "step": 5925
    },
    {
      "epoch": 2.38,
      "grad_norm": 9.803306579589844,
      "learning_rate": 3.7654320987654325e-06,
      "loss": 0.5866,
      "step": 5950
    },
    {
      "epoch": 2.39,
      "grad_norm": 10.597114562988281,
      "learning_rate": 3.734567901234568e-06,
      "loss": 0.6058,
      "step": 5975
    },
    {
      "epoch": 2.4,
      "grad_norm": 11.030874252319336,
      "learning_rate": 3.7037037037037037e-06,
      "loss": 0.6211,
      "step": 6000
    },
    {
      "epoch": 2.4,
      "eval_cer": 49.85901151405969,
      "eval_loss": 0.9176779985427856,
      "eval_runtime": 1772.9346,
      "eval_samples_per_second": 2.22,
      "eval_steps_per_second": 0.278,
      "step": 6000
    }
  ],
  "logging_steps": 25,
  "max_steps": 9000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 1000,
  "total_flos": 2.770419843072e+19,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}