whisper-small-vi / trainer_state.json
tuananh7198's picture
End of training
bc55954
raw
history blame
5.71 kB
{
"best_metric": 25.992542224171967,
"best_model_checkpoint": "./checkpoint-1000",
"epoch": 62.008,
"global_step": 1000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 1.01,
"learning_rate": 2.3000000000000004e-06,
"loss": 1.9612,
"step": 25
},
{
"epoch": 3.0,
"learning_rate": 4.800000000000001e-06,
"loss": 0.6912,
"step": 50
},
{
"epoch": 4.01,
"learning_rate": 7.3e-06,
"loss": 0.2219,
"step": 75
},
{
"epoch": 6.0,
"learning_rate": 9.800000000000001e-06,
"loss": 0.0839,
"step": 100
},
{
"epoch": 7.01,
"learning_rate": 9.744444444444445e-06,
"loss": 0.0266,
"step": 125
},
{
"epoch": 9.01,
"learning_rate": 9.466666666666667e-06,
"loss": 0.0098,
"step": 150
},
{
"epoch": 10.02,
"learning_rate": 9.188888888888889e-06,
"loss": 0.0058,
"step": 175
},
{
"epoch": 12.01,
"learning_rate": 8.91111111111111e-06,
"loss": 0.0025,
"step": 200
},
{
"epoch": 14.0,
"learning_rate": 8.633333333333334e-06,
"loss": 0.0017,
"step": 225
},
{
"epoch": 15.01,
"learning_rate": 8.355555555555556e-06,
"loss": 0.0013,
"step": 250
},
{
"epoch": 17.0,
"learning_rate": 8.077777777777778e-06,
"loss": 0.0011,
"step": 275
},
{
"epoch": 18.01,
"learning_rate": 7.800000000000002e-06,
"loss": 0.001,
"step": 300
},
{
"epoch": 20.0,
"learning_rate": 7.5222222222222226e-06,
"loss": 0.0008,
"step": 325
},
{
"epoch": 21.01,
"learning_rate": 7.244444444444445e-06,
"loss": 0.0007,
"step": 350
},
{
"epoch": 23.01,
"learning_rate": 6.966666666666667e-06,
"loss": 0.0007,
"step": 375
},
{
"epoch": 24.02,
"learning_rate": 6.688888888888889e-06,
"loss": 0.0006,
"step": 400
},
{
"epoch": 26.01,
"learning_rate": 6.411111111111111e-06,
"loss": 0.0006,
"step": 425
},
{
"epoch": 28.0,
"learning_rate": 6.133333333333334e-06,
"loss": 0.0005,
"step": 450
},
{
"epoch": 29.01,
"learning_rate": 5.855555555555556e-06,
"loss": 0.0005,
"step": 475
},
{
"epoch": 31.0,
"learning_rate": 5.577777777777778e-06,
"loss": 0.0005,
"step": 500
},
{
"epoch": 32.01,
"learning_rate": 5.300000000000001e-06,
"loss": 0.0004,
"step": 525
},
{
"epoch": 34.01,
"learning_rate": 5.022222222222223e-06,
"loss": 0.0004,
"step": 550
},
{
"epoch": 35.02,
"learning_rate": 4.744444444444445e-06,
"loss": 0.0004,
"step": 575
},
{
"epoch": 37.01,
"learning_rate": 4.4666666666666665e-06,
"loss": 0.0004,
"step": 600
},
{
"epoch": 39.0,
"learning_rate": 4.188888888888889e-06,
"loss": 0.0004,
"step": 625
},
{
"epoch": 40.01,
"learning_rate": 3.911111111111112e-06,
"loss": 0.0004,
"step": 650
},
{
"epoch": 42.0,
"learning_rate": 3.633333333333334e-06,
"loss": 0.0003,
"step": 675
},
{
"epoch": 43.01,
"learning_rate": 3.3555555555555557e-06,
"loss": 0.0003,
"step": 700
},
{
"epoch": 45.01,
"learning_rate": 3.077777777777778e-06,
"loss": 0.0003,
"step": 725
},
{
"epoch": 46.01,
"learning_rate": 2.8000000000000003e-06,
"loss": 0.0003,
"step": 750
},
{
"epoch": 48.01,
"learning_rate": 2.5222222222222226e-06,
"loss": 0.0003,
"step": 775
},
{
"epoch": 49.02,
"learning_rate": 2.2444444444444445e-06,
"loss": 0.0003,
"step": 800
},
{
"epoch": 51.01,
"learning_rate": 1.9666666666666668e-06,
"loss": 0.0003,
"step": 825
},
{
"epoch": 53.0,
"learning_rate": 1.688888888888889e-06,
"loss": 0.0003,
"step": 850
},
{
"epoch": 54.01,
"learning_rate": 1.4111111111111111e-06,
"loss": 0.0003,
"step": 875
},
{
"epoch": 56.0,
"learning_rate": 1.1333333333333334e-06,
"loss": 0.0003,
"step": 900
},
{
"epoch": 57.01,
"learning_rate": 8.555555555555556e-07,
"loss": 0.0003,
"step": 925
},
{
"epoch": 59.01,
"learning_rate": 5.777777777777778e-07,
"loss": 0.0003,
"step": 950
},
{
"epoch": 60.02,
"learning_rate": 3.0000000000000004e-07,
"loss": 0.0003,
"step": 975
},
{
"epoch": 62.01,
"learning_rate": 2.2222222222222224e-08,
"loss": 0.0003,
"step": 1000
},
{
"epoch": 62.01,
"eval_loss": 0.7276960611343384,
"eval_runtime": 439.4606,
"eval_samples_per_second": 2.815,
"eval_steps_per_second": 0.046,
"eval_wer": 25.992542224171967,
"step": 1000
},
{
"epoch": 62.01,
"step": 1000,
"total_flos": 8.94845609312256e+18,
"train_loss": 0.07548951799841598,
"train_runtime": 17725.7145,
"train_samples_per_second": 14.442,
"train_steps_per_second": 0.056
}
],
"max_steps": 1000,
"num_train_epochs": 9223372036854775807,
"total_flos": 8.94845609312256e+18,
"trial_name": null,
"trial_params": null
}