wav2vec2-large-xlsr-53-ja / trainer_state.json
vutankiet2901's picture
End of training
5889a8c
raw
history blame
4.71 kB
{
"best_metric": 0.42136871814727783,
"best_model_checkpoint": "wav2vec2-large-xlsr-53-ja/checkpoint-15000",
"epoch": 49.99842519685039,
"global_step": 15850,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 4.73,
"learning_rate": 0.00022439999999999998,
"loss": 4.7776,
"step": 1500
},
{
"epoch": 4.73,
"eval_cer": 0.8489376326091504,
"eval_loss": 2.95402193069458,
"eval_runtime": 195.078,
"eval_samples_per_second": 22.914,
"eval_steps_per_second": 2.866,
"eval_wer": 0.9772047095930181,
"step": 1500
},
{
"epoch": 9.46,
"learning_rate": 0.00027842599277978335,
"loss": 1.9076,
"step": 3000
},
{
"epoch": 9.46,
"eval_cer": 0.24839972506947972,
"eval_loss": 0.7145683169364929,
"eval_runtime": 197.8842,
"eval_samples_per_second": 22.589,
"eval_steps_per_second": 2.825,
"eval_wer": 0.5371229906274081,
"step": 3000
},
{
"epoch": 14.2,
"learning_rate": 0.0002459783393501805,
"loss": 1.507,
"step": 4500
},
{
"epoch": 14.2,
"eval_cer": 0.21956190419269042,
"eval_loss": 0.5843430757522583,
"eval_runtime": 196.1707,
"eval_samples_per_second": 22.786,
"eval_steps_per_second": 2.85,
"eval_wer": 0.4689342484901704,
"step": 4500
},
{
"epoch": 18.93,
"learning_rate": 0.0002135090252707581,
"loss": 1.3742,
"step": 6000
},
{
"epoch": 18.93,
"eval_cer": 0.19882855691360607,
"eval_loss": 0.5286233425140381,
"eval_runtime": 195.147,
"eval_samples_per_second": 22.906,
"eval_steps_per_second": 2.865,
"eval_wer": 0.4321248723141162,
"step": 6000
},
{
"epoch": 23.66,
"learning_rate": 0.00018101805054151622,
"loss": 1.2776,
"step": 7500
},
{
"epoch": 23.66,
"eval_cer": 0.1869527537877656,
"eval_loss": 0.5006501078605652,
"eval_runtime": 193.294,
"eval_samples_per_second": 23.125,
"eval_steps_per_second": 2.892,
"eval_wer": 0.4055841293166789,
"step": 7500
},
{
"epoch": 28.39,
"learning_rate": 0.00014857039711191334,
"loss": 1.2003,
"step": 9000
},
{
"epoch": 28.39,
"eval_cer": 0.18024683979320444,
"eval_loss": 0.46759462356567383,
"eval_runtime": 194.4872,
"eval_samples_per_second": 22.984,
"eval_steps_per_second": 2.874,
"eval_wer": 0.3848138922241537,
"step": 9000
},
{
"epoch": 33.12,
"learning_rate": 0.00011607942238267148,
"loss": 1.1281,
"step": 10500
},
{
"epoch": 33.12,
"eval_cer": 0.17200490093536144,
"eval_loss": 0.4524390697479248,
"eval_runtime": 194.4065,
"eval_samples_per_second": 22.993,
"eval_steps_per_second": 2.875,
"eval_wer": 0.3694199028691242,
"step": 10500
},
{
"epoch": 37.85,
"learning_rate": 8.35884476534296e-05,
"loss": 1.0657,
"step": 12000
},
{
"epoch": 37.85,
"eval_cer": 0.1680662223948839,
"eval_loss": 0.44492024183273315,
"eval_runtime": 194.6254,
"eval_samples_per_second": 22.967,
"eval_steps_per_second": 2.872,
"eval_wer": 0.35898998225838247,
"step": 12000
},
{
"epoch": 42.59,
"learning_rate": 5.111913357400722e-05,
"loss": 1.0129,
"step": 13500
},
{
"epoch": 42.59,
"eval_cer": 0.16168902967456594,
"eval_loss": 0.42659756541252136,
"eval_runtime": 194.2959,
"eval_samples_per_second": 23.006,
"eval_steps_per_second": 2.877,
"eval_wer": 0.34232361427214564,
"step": 13500
},
{
"epoch": 47.32,
"learning_rate": 1.8649819494584836e-05,
"loss": 0.9691,
"step": 15000
},
{
"epoch": 47.32,
"eval_cer": 0.1587305382063772,
"eval_loss": 0.42136871814727783,
"eval_runtime": 195.4013,
"eval_samples_per_second": 22.876,
"eval_steps_per_second": 2.861,
"eval_wer": 0.33750291213419115,
"step": 15000
},
{
"epoch": 50.0,
"step": 15850,
"total_flos": 7.595274391472413e+19,
"train_loss": 1.585643256635696,
"train_runtime": 29809.7185,
"train_samples_per_second": 17.035,
"train_steps_per_second": 0.532
}
],
"max_steps": 15850,
"num_train_epochs": 50,
"total_flos": 7.595274391472413e+19,
"trial_name": null,
"trial_params": null
}