{
  "best_metric": 96.39777735198314,
  "best_model_checkpoint": "./whisper-small-ha-v3/checkpoint-1000",
  "epoch": 12.738853503184714,
  "eval_steps": 500,
  "global_step": 2000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1592356687898089,
      "grad_norm": 202.8230438232422,
      "learning_rate": 0.0002,
      "loss": 3.0913,
      "step": 25
    },
    {
      "epoch": 0.3184713375796178,
      "grad_norm": 20.680511474609375,
      "learning_rate": 0.00045000000000000004,
      "loss": 2.931,
      "step": 50
    },
    {
      "epoch": 0.47770700636942676,
      "grad_norm": 47.82948303222656,
      "learning_rate": 0.0005,
      "loss": 3.5464,
      "step": 75
    },
    {
      "epoch": 0.6369426751592356,
      "grad_norm": 10.464394569396973,
      "learning_rate": 0.0005,
      "loss": 5.0146,
      "step": 100
    },
    {
      "epoch": 0.7961783439490446,
      "grad_norm": 9.31248664855957,
      "learning_rate": 0.0005,
      "loss": 3.5674,
      "step": 125
    },
    {
      "epoch": 0.9554140127388535,
      "grad_norm": 8.952727317810059,
      "learning_rate": 0.0005,
      "loss": 3.5191,
      "step": 150
    },
    {
      "epoch": 1.1146496815286624,
      "grad_norm": 8.769633293151855,
      "learning_rate": 0.0005,
      "loss": 2.9062,
      "step": 175
    },
    {
      "epoch": 1.2738853503184713,
      "grad_norm": 8.326931953430176,
      "learning_rate": 0.0005,
      "loss": 2.7467,
      "step": 200
    },
    {
      "epoch": 1.4331210191082802,
      "grad_norm": 8.937690734863281,
      "learning_rate": 0.0005,
      "loss": 2.8934,
      "step": 225
    },
    {
      "epoch": 1.5923566878980893,
      "grad_norm": 7.839733600616455,
      "learning_rate": 0.0005,
      "loss": 2.9366,
      "step": 250
    },
    {
      "epoch": 1.7515923566878981,
      "grad_norm": 7.158330917358398,
      "learning_rate": 0.0005,
      "loss": 2.8352,
      "step": 275
    },
    {
      "epoch": 1.910828025477707,
      "grad_norm": 7.30534029006958,
      "learning_rate": 0.0005,
      "loss": 2.9709,
      "step": 300
    },
    {
      "epoch": 2.070063694267516,
      "grad_norm": 7.713191509246826,
      "learning_rate": 0.0005,
      "loss": 2.4781,
      "step": 325
    },
    {
      "epoch": 2.229299363057325,
      "grad_norm": 7.518861293792725,
      "learning_rate": 0.0005,
      "loss": 2.0545,
      "step": 350
    },
    {
      "epoch": 2.388535031847134,
      "grad_norm": 8.62714958190918,
      "learning_rate": 0.0005,
      "loss": 2.1755,
      "step": 375
    },
    {
      "epoch": 2.5477707006369426,
      "grad_norm": 8.683177947998047,
      "learning_rate": 0.0005,
      "loss": 2.2715,
      "step": 400
    },
    {
      "epoch": 2.7070063694267517,
      "grad_norm": 7.4501142501831055,
      "learning_rate": 0.0005,
      "loss": 2.3387,
      "step": 425
    },
    {
      "epoch": 2.8662420382165603,
      "grad_norm": 8.682276725769043,
      "learning_rate": 0.0005,
      "loss": 2.3904,
      "step": 450
    },
    {
      "epoch": 3.0254777070063694,
      "grad_norm": 7.98887825012207,
      "learning_rate": 0.0005,
      "loss": 2.192,
      "step": 475
    },
    {
      "epoch": 3.1847133757961785,
      "grad_norm": 8.835172653198242,
      "learning_rate": 0.0005,
      "loss": 1.5826,
      "step": 500
    },
    {
      "epoch": 3.1847133757961785,
      "eval_loss": 3.9010848999023438,
      "eval_runtime": 343.8775,
      "eval_samples_per_second": 1.919,
      "eval_steps_per_second": 0.122,
      "eval_wer": 180.03448936577888,
      "eval_wer_ortho": 183.0078125,
      "step": 500
    },
    {
      "epoch": 3.343949044585987,
      "grad_norm": 7.396618843078613,
      "learning_rate": 0.0005,
      "loss": 1.6942,
      "step": 525
    },
    {
      "epoch": 3.5031847133757963,
      "grad_norm": 7.276708602905273,
      "learning_rate": 0.0005,
      "loss": 1.7758,
      "step": 550
    },
    {
      "epoch": 3.662420382165605,
      "grad_norm": 8.082304000854492,
      "learning_rate": 0.0005,
      "loss": 1.8561,
      "step": 575
    },
    {
      "epoch": 3.821656050955414,
      "grad_norm": 7.295908451080322,
      "learning_rate": 0.0005,
      "loss": 1.9064,
      "step": 600
    },
    {
      "epoch": 3.980891719745223,
      "grad_norm": 7.9326491355896,
      "learning_rate": 0.0005,
      "loss": 1.9251,
      "step": 625
    },
    {
      "epoch": 4.140127388535032,
      "grad_norm": 7.216257095336914,
      "learning_rate": 0.0005,
      "loss": 1.3606,
      "step": 650
    },
    {
      "epoch": 4.2993630573248405,
      "grad_norm": 6.987307548522949,
      "learning_rate": 0.0005,
      "loss": 1.382,
      "step": 675
    },
    {
      "epoch": 4.45859872611465,
      "grad_norm": 8.181788444519043,
      "learning_rate": 0.0005,
      "loss": 1.4235,
      "step": 700
    },
    {
      "epoch": 4.617834394904459,
      "grad_norm": 7.226937294006348,
      "learning_rate": 0.0005,
      "loss": 1.5352,
      "step": 725
    },
    {
      "epoch": 4.777070063694268,
      "grad_norm": 7.664785385131836,
      "learning_rate": 0.0005,
      "loss": 1.5803,
      "step": 750
    },
    {
      "epoch": 4.936305732484076,
      "grad_norm": 8.353466987609863,
      "learning_rate": 0.0005,
      "loss": 1.6793,
      "step": 775
    },
    {
      "epoch": 5.095541401273885,
      "grad_norm": 7.170167922973633,
      "learning_rate": 0.0005,
      "loss": 1.3162,
      "step": 800
    },
    {
      "epoch": 5.254777070063694,
      "grad_norm": 7.019118309020996,
      "learning_rate": 0.0005,
      "loss": 1.1988,
      "step": 825
    },
    {
      "epoch": 5.414012738853503,
      "grad_norm": 6.39375638961792,
      "learning_rate": 0.0005,
      "loss": 1.2476,
      "step": 850
    },
    {
      "epoch": 5.573248407643312,
      "grad_norm": 7.22137451171875,
      "learning_rate": 0.0005,
      "loss": 1.325,
      "step": 875
    },
    {
      "epoch": 5.732484076433121,
      "grad_norm": 6.7961883544921875,
      "learning_rate": 0.0005,
      "loss": 1.3379,
      "step": 900
    },
    {
      "epoch": 5.89171974522293,
      "grad_norm": 7.7992377281188965,
      "learning_rate": 0.0005,
      "loss": 1.3845,
      "step": 925
    },
    {
      "epoch": 6.050955414012739,
      "grad_norm": 6.209515571594238,
      "learning_rate": 0.0005,
      "loss": 1.2692,
      "step": 950
    },
    {
      "epoch": 6.210191082802548,
      "grad_norm": 6.86682653427124,
      "learning_rate": 0.0005,
      "loss": 1.1113,
      "step": 975
    },
    {
      "epoch": 6.369426751592357,
      "grad_norm": 6.647078037261963,
      "learning_rate": 0.0005,
      "loss": 1.1378,
      "step": 1000
    },
    {
      "epoch": 6.369426751592357,
      "eval_loss": 4.400506019592285,
      "eval_runtime": 280.0441,
      "eval_samples_per_second": 2.357,
      "eval_steps_per_second": 0.15,
      "eval_wer": 96.39777735198314,
      "eval_wer_ortho": 96.640625,
      "step": 1000
    },
    {
      "epoch": 6.528662420382165,
      "grad_norm": 7.238087177276611,
      "learning_rate": 0.0005,
      "loss": 1.2028,
      "step": 1025
    },
    {
      "epoch": 6.687898089171974,
      "grad_norm": 6.603641033172607,
      "learning_rate": 0.0005,
      "loss": 1.2026,
      "step": 1050
    },
    {
      "epoch": 6.8471337579617835,
      "grad_norm": 6.680185794830322,
      "learning_rate": 0.0005,
      "loss": 1.2335,
      "step": 1075
    },
    {
      "epoch": 7.006369426751593,
      "grad_norm": 5.673605442047119,
      "learning_rate": 0.0005,
      "loss": 1.2663,
      "step": 1100
    },
    {
      "epoch": 7.165605095541402,
      "grad_norm": 6.584240913391113,
      "learning_rate": 0.0005,
      "loss": 0.9781,
      "step": 1125
    },
    {
      "epoch": 7.32484076433121,
      "grad_norm": 6.363912582397461,
      "learning_rate": 0.0005,
      "loss": 1.04,
      "step": 1150
    },
    {
      "epoch": 7.484076433121019,
      "grad_norm": 5.395627021789551,
      "learning_rate": 0.0005,
      "loss": 1.1184,
      "step": 1175
    },
    {
      "epoch": 7.643312101910828,
      "grad_norm": 6.268621921539307,
      "learning_rate": 0.0005,
      "loss": 1.1184,
      "step": 1200
    },
    {
      "epoch": 7.802547770700637,
      "grad_norm": 6.605058193206787,
      "learning_rate": 0.0005,
      "loss": 1.1101,
      "step": 1225
    },
    {
      "epoch": 7.961783439490446,
      "grad_norm": 6.293459892272949,
      "learning_rate": 0.0005,
      "loss": 1.1396,
      "step": 1250
    },
    {
      "epoch": 8.121019108280255,
      "grad_norm": 6.258079528808594,
      "learning_rate": 0.0005,
      "loss": 0.9623,
      "step": 1275
    },
    {
      "epoch": 8.280254777070065,
      "grad_norm": 5.724878787994385,
      "learning_rate": 0.0005,
      "loss": 0.9693,
      "step": 1300
    },
    {
      "epoch": 8.439490445859873,
      "grad_norm": 5.03961181640625,
      "learning_rate": 0.0005,
      "loss": 0.9993,
      "step": 1325
    },
    {
      "epoch": 8.598726114649681,
      "grad_norm": 6.147229194641113,
      "learning_rate": 0.0005,
      "loss": 1.0598,
      "step": 1350
    },
    {
      "epoch": 8.757961783439491,
      "grad_norm": 7.066701889038086,
      "learning_rate": 0.0005,
      "loss": 1.0835,
      "step": 1375
    },
    {
      "epoch": 8.9171974522293,
      "grad_norm": 5.424177646636963,
      "learning_rate": 0.0005,
      "loss": 1.1042,
      "step": 1400
    },
    {
      "epoch": 9.07643312101911,
      "grad_norm": 5.849576473236084,
      "learning_rate": 0.0005,
      "loss": 0.9797,
      "step": 1425
    },
    {
      "epoch": 9.235668789808917,
      "grad_norm": 4.9155120849609375,
      "learning_rate": 0.0005,
      "loss": 0.8813,
      "step": 1450
    },
    {
      "epoch": 9.394904458598726,
      "grad_norm": 6.510490417480469,
      "learning_rate": 0.0005,
      "loss": 0.9536,
      "step": 1475
    },
    {
      "epoch": 9.554140127388536,
      "grad_norm": 5.549797058105469,
      "learning_rate": 0.0005,
      "loss": 0.9735,
      "step": 1500
    },
    {
      "epoch": 9.554140127388536,
      "eval_loss": 4.821861267089844,
      "eval_runtime": 286.5747,
      "eval_samples_per_second": 2.303,
      "eval_steps_per_second": 0.147,
      "eval_wer": 103.04656064380148,
      "eval_wer_ortho": 104.1015625,
      "step": 1500
    },
    {
      "epoch": 9.713375796178344,
      "grad_norm": 5.422865390777588,
      "learning_rate": 0.0005,
      "loss": 1.0079,
      "step": 1525
    },
    {
      "epoch": 9.872611464968152,
      "grad_norm": 5.475556373596191,
      "learning_rate": 0.0005,
      "loss": 1.0126,
      "step": 1550
    },
    {
      "epoch": 10.031847133757962,
      "grad_norm": 5.356685161590576,
      "learning_rate": 0.0005,
      "loss": 0.9727,
      "step": 1575
    },
    {
      "epoch": 10.19108280254777,
      "grad_norm": 5.031153202056885,
      "learning_rate": 0.0005,
      "loss": 0.8431,
      "step": 1600
    },
    {
      "epoch": 10.35031847133758,
      "grad_norm": 5.827383518218994,
      "learning_rate": 0.0005,
      "loss": 0.8888,
      "step": 1625
    },
    {
      "epoch": 10.509554140127388,
      "grad_norm": 5.030758857727051,
      "learning_rate": 0.0005,
      "loss": 0.9165,
      "step": 1650
    },
    {
      "epoch": 10.668789808917197,
      "grad_norm": 5.023013114929199,
      "learning_rate": 0.0005,
      "loss": 0.9541,
      "step": 1675
    },
    {
      "epoch": 10.828025477707007,
      "grad_norm": 5.8825602531433105,
      "learning_rate": 0.0005,
      "loss": 0.9576,
      "step": 1700
    },
    {
      "epoch": 10.987261146496815,
      "grad_norm": 5.114201068878174,
      "learning_rate": 0.0005,
      "loss": 0.9793,
      "step": 1725
    },
    {
      "epoch": 11.146496815286625,
      "grad_norm": 4.3722333908081055,
      "learning_rate": 0.0005,
      "loss": 0.7909,
      "step": 1750
    },
    {
      "epoch": 11.305732484076433,
      "grad_norm": 4.9860382080078125,
      "learning_rate": 0.0005,
      "loss": 0.858,
      "step": 1775
    },
    {
      "epoch": 11.464968152866241,
      "grad_norm": 5.144904613494873,
      "learning_rate": 0.0005,
      "loss": 0.8842,
      "step": 1800
    },
    {
      "epoch": 11.624203821656051,
      "grad_norm": 4.30189847946167,
      "learning_rate": 0.0005,
      "loss": 0.8925,
      "step": 1825
    },
    {
      "epoch": 11.78343949044586,
      "grad_norm": 5.091893672943115,
      "learning_rate": 0.0005,
      "loss": 0.9188,
      "step": 1850
    },
    {
      "epoch": 11.94267515923567,
      "grad_norm": 5.179553508758545,
      "learning_rate": 0.0005,
      "loss": 0.9309,
      "step": 1875
    },
    {
      "epoch": 12.101910828025478,
      "grad_norm": 5.5085225105285645,
      "learning_rate": 0.0005,
      "loss": 0.8213,
      "step": 1900
    },
    {
      "epoch": 12.261146496815286,
      "grad_norm": 5.253794193267822,
      "learning_rate": 0.0005,
      "loss": 0.8155,
      "step": 1925
    },
    {
      "epoch": 12.420382165605096,
      "grad_norm": 4.998741149902344,
      "learning_rate": 0.0005,
      "loss": 0.8479,
      "step": 1950
    },
    {
      "epoch": 12.579617834394904,
      "grad_norm": 5.674376010894775,
      "learning_rate": 0.0005,
      "loss": 0.8773,
      "step": 1975
    },
    {
      "epoch": 12.738853503184714,
      "grad_norm": 5.010542869567871,
      "learning_rate": 0.0005,
      "loss": 0.8844,
      "step": 2000
    },
    {
      "epoch": 12.738853503184714,
      "eval_loss": 5.002076148986816,
      "eval_runtime": 288.6571,
      "eval_samples_per_second": 2.286,
      "eval_steps_per_second": 0.146,
      "eval_wer": 109.04387813757425,
      "eval_wer_ortho": 108.59375,
      "step": 2000
    }
  ],
  "logging_steps": 25,
  "max_steps": 5652,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 36,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 9.22088071102464e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}