wav2vec2-xls-r-common_voice-tr-ft-100sh

作者 anton-l · 下载 2 · 点赞 0 · 更新 2023-10-31

automatic-speech-recognition common_voice generated_from_trainer

wav2vec2-xls-r-common_voice-tr-ft

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the COMMON_VOICE - TR dataset. It achieves the following results on the evaluation set:

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

Training results

Training Loss Epoch Step Validation Loss Wer Cer
0.5369 17.0 500 0.6021 0.6366 0.1727
0.3542 34.0 1000 0.5265 0.4906 0.1278
0.1866 51.0 1500 0.5805 0.4768 0.1261
0.1674 68.01 2000 0.5336 0.4518 0.1186
0.19 86.0 2500 0.5676 0.4427 0.1151
0.0815 103.0 3000 0.5510 0.4268 0.1125
0.0545 120.0 3500 0.5608 0.4175 0.1099
0.0299 137.01 4000 0.5875 0.4222 0.1124
0.0267 155.0 4500 0.5882 0.4026 0.1063
0.025 172.0 5000 0.5806 0.3998 0.1053

Framework versions