trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 30.0,
  "global_step": 3870,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.78,
      "eval_loss": 4.564502716064453,
      "eval_runtime": 179.128,
      "eval_samples_per_second": 20.142,
      "eval_steps_per_second": 2.518,
      "eval_wer": 1.0,
      "step": 100
    },
    {
      "epoch": 1.55,
      "eval_loss": 2.9015955924987793,
      "eval_runtime": 175.3683,
      "eval_samples_per_second": 20.574,
      "eval_steps_per_second": 2.572,
      "eval_wer": 1.0,
      "step": 200
    },
    {
      "epoch": 2.33,
      "eval_loss": 2.266641616821289,
      "eval_runtime": 174.1906,
      "eval_samples_per_second": 20.713,
      "eval_steps_per_second": 2.589,
      "eval_wer": 1.098215533709174,
      "step": 300
    },
    {
      "epoch": 3.1,
      "eval_loss": 0.6078956127166748,
      "eval_runtime": 181.241,
      "eval_samples_per_second": 19.907,
      "eval_steps_per_second": 2.488,
      "eval_wer": 0.6375623165447515,
      "step": 400
    },
    {
      "epoch": 3.88,
      "learning_rate": 0.0002988,
      "loss": 3.2188,
      "step": 500
    },
    {
      "epoch": 3.88,
      "eval_loss": 0.49847397208213806,
      "eval_runtime": 175.9858,
      "eval_samples_per_second": 20.502,
      "eval_steps_per_second": 2.563,
      "eval_wer": 0.5007687648511392,
      "step": 500
    },
    {
      "epoch": 4.65,
      "eval_loss": 0.447698175907135,
      "eval_runtime": 175.6036,
      "eval_samples_per_second": 20.546,
      "eval_steps_per_second": 2.568,
      "eval_wer": 0.44690863346223736,
      "step": 600
    },
    {
      "epoch": 5.43,
      "eval_loss": 0.39529213309288025,
      "eval_runtime": 175.4478,
      "eval_samples_per_second": 20.565,
      "eval_steps_per_second": 2.571,
      "eval_wer": 0.3914643805618972,
      "step": 700
    },
    {
      "epoch": 6.2,
      "eval_loss": 0.43193477392196655,
      "eval_runtime": 176.2224,
      "eval_samples_per_second": 20.474,
      "eval_steps_per_second": 2.559,
      "eval_wer": 0.3921166658901365,
      "step": 800
    },
    {
      "epoch": 6.98,
      "eval_loss": 0.4170827865600586,
      "eval_runtime": 175.9073,
      "eval_samples_per_second": 20.511,
      "eval_steps_per_second": 2.564,
      "eval_wer": 0.3698457811116806,
      "step": 900
    },
    {
      "epoch": 7.75,
      "learning_rate": 0.00025566765578635014,
      "loss": 0.2193,
      "step": 1000
    },
    {
      "epoch": 7.75,
      "eval_loss": 0.3956995904445648,
      "eval_runtime": 178.52,
      "eval_samples_per_second": 20.211,
      "eval_steps_per_second": 2.526,
      "eval_wer": 0.36001490937893116,
      "step": 1000
    },
    {
      "epoch": 8.53,
      "eval_loss": 0.37300992012023926,
      "eval_runtime": 175.8463,
      "eval_samples_per_second": 20.518,
      "eval_steps_per_second": 2.565,
      "eval_wer": 0.34929879327214275,
      "step": 1100
    },
    {
      "epoch": 9.3,
      "eval_loss": 0.3779752850532532,
      "eval_runtime": 176.2153,
      "eval_samples_per_second": 20.475,
      "eval_steps_per_second": 2.559,
      "eval_wer": 0.3348087406233984,
      "step": 1200
    },
    {
      "epoch": 10.08,
      "eval_loss": 0.41326919198036194,
      "eval_runtime": 173.7883,
      "eval_samples_per_second": 20.761,
      "eval_steps_per_second": 2.595,
      "eval_wer": 0.35680007454689466,
      "step": 1300
    },
    {
      "epoch": 10.85,
      "eval_loss": 0.3984449505805969,
      "eval_runtime": 175.4199,
      "eval_samples_per_second": 20.568,
      "eval_steps_per_second": 2.571,
      "eval_wer": 0.31929366817313515,
      "step": 1400
    },
    {
      "epoch": 11.63,
      "learning_rate": 0.00021115727002967357,
      "loss": 0.1129,
      "step": 1500
    },
    {
      "epoch": 11.63,
      "eval_loss": 0.38447898626327515,
      "eval_runtime": 175.2875,
      "eval_samples_per_second": 20.583,
      "eval_steps_per_second": 2.573,
      "eval_wer": 0.3174299958067372,
      "step": 1500
    },
    {
      "epoch": 12.4,
      "eval_loss": 0.3882218301296234,
      "eval_runtime": 175.3702,
      "eval_samples_per_second": 20.574,
      "eval_steps_per_second": 2.572,
      "eval_wer": 0.3162186087685785,
      "step": 1600
    },
    {
      "epoch": 13.18,
      "eval_loss": 0.39824405312538147,
      "eval_runtime": 175.3906,
      "eval_samples_per_second": 20.571,
      "eval_steps_per_second": 2.571,
      "eval_wer": 0.3008433117457951,
      "step": 1700
    },
    {
      "epoch": 13.95,
      "eval_loss": 0.3901589810848236,
      "eval_runtime": 177.0061,
      "eval_samples_per_second": 20.383,
      "eval_steps_per_second": 2.548,
      "eval_wer": 0.3198061780738946,
      "step": 1800
    },
    {
      "epoch": 14.73,
      "eval_loss": 0.4082184135913849,
      "eval_runtime": 175.0779,
      "eval_samples_per_second": 20.608,
      "eval_steps_per_second": 2.576,
      "eval_wer": 0.3237198900433304,
      "step": 1900
    },
    {
      "epoch": 15.5,
      "learning_rate": 0.000166646884272997,
      "loss": 0.0765,
      "step": 2000
    },
    {
      "epoch": 15.5,
      "eval_loss": 0.3732178211212158,
      "eval_runtime": 175.4721,
      "eval_samples_per_second": 20.562,
      "eval_steps_per_second": 2.57,
      "eval_wer": 0.3125844476541024,
      "step": 2000
    },
    {
      "epoch": 16.28,
      "eval_loss": 0.3892667293548584,
      "eval_runtime": 178.0744,
      "eval_samples_per_second": 20.261,
      "eval_steps_per_second": 2.533,
      "eval_wer": 0.3000978427992359,
      "step": 2100
    },
    {
      "epoch": 17.05,
      "eval_loss": 0.4168277978897095,
      "eval_runtime": 176.2583,
      "eval_samples_per_second": 20.47,
      "eval_steps_per_second": 2.559,
      "eval_wer": 0.308344593020547,
      "step": 2200
    },
    {
      "epoch": 17.83,
      "eval_loss": 0.4192778170108795,
      "eval_runtime": 172.3953,
      "eval_samples_per_second": 20.929,
      "eval_steps_per_second": 2.616,
      "eval_wer": 0.3044308810511112,
      "step": 2300
    },
    {
      "epoch": 18.6,
      "eval_loss": 0.40058156847953796,
      "eval_runtime": 174.3893,
      "eval_samples_per_second": 20.689,
      "eval_steps_per_second": 2.586,
      "eval_wer": 0.3013092298373946,
      "step": 2400
    },
    {
      "epoch": 19.38,
      "learning_rate": 0.00012213649851632047,
      "loss": 0.0588,
      "step": 2500
    },
    {
      "epoch": 19.38,
      "eval_loss": 0.38357821106910706,
      "eval_runtime": 180.4646,
      "eval_samples_per_second": 19.993,
      "eval_steps_per_second": 2.499,
      "eval_wer": 0.2892419512649676,
      "step": 2500
    },
    {
      "epoch": 20.16,
      "eval_loss": 0.3760845959186554,
      "eval_runtime": 191.865,
      "eval_samples_per_second": 18.805,
      "eval_steps_per_second": 2.351,
      "eval_wer": 0.2902669710664865,
      "step": 2600
    },
    {
      "epoch": 20.93,
      "eval_loss": 0.38948509097099304,
      "eval_runtime": 181.0859,
      "eval_samples_per_second": 19.924,
      "eval_steps_per_second": 2.491,
      "eval_wer": 0.29301588780692356,
      "step": 2700
    },
    {
      "epoch": 21.71,
      "eval_loss": 0.3884966969490051,
      "eval_runtime": 179.119,
      "eval_samples_per_second": 20.143,
      "eval_steps_per_second": 2.518,
      "eval_wer": 0.27913152867725854,
      "step": 2800
    },
    {
      "epoch": 22.48,
      "eval_loss": 0.3901614248752594,
      "eval_runtime": 182.8151,
      "eval_samples_per_second": 19.736,
      "eval_steps_per_second": 2.467,
      "eval_wer": 0.2891487676466477,
      "step": 2900
    },
    {
      "epoch": 23.26,
      "learning_rate": 7.762611275964391e-05,
      "loss": 0.0448,
      "step": 3000
    },
    {
      "epoch": 23.26,
      "eval_loss": 0.42003825306892395,
      "eval_runtime": 180.6554,
      "eval_samples_per_second": 19.972,
      "eval_steps_per_second": 2.496,
      "eval_wer": 0.2849089130130923,
      "step": 3000
    },
    {
      "epoch": 24.03,
      "eval_loss": 0.40127792954444885,
      "eval_runtime": 180.093,
      "eval_samples_per_second": 20.034,
      "eval_steps_per_second": 2.504,
      "eval_wer": 0.27987699762381774,
      "step": 3100
    },
    {
      "epoch": 24.81,
      "eval_loss": 0.4039434492588043,
      "eval_runtime": 182.9792,
      "eval_samples_per_second": 19.718,
      "eval_steps_per_second": 2.465,
      "eval_wer": 0.273121185295625,
      "step": 3200
    },
    {
      "epoch": 25.58,
      "eval_loss": 0.397048681974411,
      "eval_runtime": 186.4145,
      "eval_samples_per_second": 19.355,
      "eval_steps_per_second": 2.419,
      "eval_wer": 0.26473465964683407,
      "step": 3300
    },
    {
      "epoch": 26.36,
      "eval_loss": 0.4080738127231598,
      "eval_runtime": 183.3845,
      "eval_samples_per_second": 19.675,
      "eval_steps_per_second": 2.459,
      "eval_wer": 0.26902110608954943,
      "step": 3400
    },
    {
      "epoch": 27.13,
      "learning_rate": 3.311572700296736e-05,
      "loss": 0.0351,
      "step": 3500
    },
    {
      "epoch": 27.13,
      "eval_loss": 0.4090190827846527,
      "eval_runtime": 188.4549,
      "eval_samples_per_second": 19.145,
      "eval_steps_per_second": 2.393,
      "eval_wer": 0.26743698457811116,
      "step": 3500
    },
    {
      "epoch": 27.91,
      "eval_loss": 0.3952561318874359,
      "eval_runtime": 183.4934,
      "eval_samples_per_second": 19.663,
      "eval_steps_per_second": 2.458,
      "eval_wer": 0.26627218934911245,
      "step": 3600
    },
    {
      "epoch": 28.68,
      "eval_loss": 0.40437063574790955,
      "eval_runtime": 188.3196,
      "eval_samples_per_second": 19.159,
      "eval_steps_per_second": 2.395,
      "eval_wer": 0.26496761869263386,
      "step": 3700
    },
    {
      "epoch": 29.46,
      "eval_loss": 0.3968600630760193,
      "eval_runtime": 182.2466,
      "eval_samples_per_second": 19.797,
      "eval_steps_per_second": 2.475,
      "eval_wer": 0.26459488421935423,
      "step": 3800
    },
    {
      "epoch": 30.0,
      "step": 3870,
      "total_flos": 4.209827274605221e+19,
      "train_loss": 0.4894287838800317,
      "train_runtime": 22523.7187,
      "train_samples_per_second": 10.974,
      "train_steps_per_second": 0.172
    }
  ],
  "max_steps": 3870,
  "num_train_epochs": 30,
  "total_flos": 4.209827274605221e+19,
  "trial_name": null,
  "trial_params": null
}