Hui Lu

> Thanks, could you point me to the repo you used to train the hi-fi gan. I would like to train it using my own dataset. Check it out here:...

> 我下载了标贝的数据集，然后预处理，然后直接用你的预训练的模型执行inference.py得到mels,然后再用你上传的Hifi-gan的模型来生成wav,得到的结果听上去没有你[https://light1726.github.io/vaenar-tts/里的sample效果好，而且我发现你的是16kHZ的，而我得到的是22.05KHZ的](https://light1726.github.io/vaenar-tts/%E9%87%8C%E7%9A%84sample%E6%95%88%E6%9E%9C%E5%A5%BD%EF%BC%8C%E8%80%8C%E4%B8%94%E6%88%91%E5%8F%91%E7%8E%B0%E4%BD%A0%E7%9A%84%E6%98%AF16kHZ%E7%9A%84%EF%BC%8C%E8%80%8C%E6%88%91%E5%BE%97%E5%88%B0%E7%9A%84%E6%98%AF22.05KHZ%E7%9A%84) CUDA_VISIBLE_DEVICES=0 TF_FORCE_GPU_ALLOW_GROWTH=true python inference.py --dataset databaker --test_dir ./pretrained/db-test-2000 --data_dir ./biaobei/tfrecords/ --batch_size 16 --write_wavs true --draw_alignments true --ckpt_path ./pretrained/db-model_dir/ckpt-2000 > > python vocoder.py -i ./mels-pre/ -o ./waves-pre/ --checkpoint_file ./model_dir/g_01160000...

> > > 我下载了标贝的数据集，然后预处理，然后直接用你的预训练的模型执行inference.py得到mels,然后再用你上传的Hifi-gan的模型来生成wav,得到的结果听上去没有你[https://light1726.github.io/vaenar-tts/里的sample效果好，而且我发现你的是16kHZ的，而我得到的是22.05KHZ的](https://light1726.github.io/vaenar-tts/%E9%87%8C%E7%9A%84sample%E6%95%88%E6%9E%9C%E5%A5%BD%EF%BC%8C%E8%80%8C%E4%B8%94%E6%88%91%E5%8F%91%E7%8E%B0%E4%BD%A0%E7%9A%84%E6%98%AF16kHZ%E7%9A%84%EF%BC%8C%E8%80%8C%E6%88%91%E5%BE%97%E5%88%B0%E7%9A%84%E6%98%AF22.05KHZ%E7%9A%84) CUDA_VISIBLE_DEVICES=0 TF_FORCE_GPU_ALLOW_GROWTH=true python inference.py --dataset databaker --test_dir ./pretrained/db-test-2000 --data_dir ./biaobei/tfrecords/ --batch_size 16 --write_wavs true --draw_alignments true --ckpt_path ./pretrained/db-model_dir/ckpt-2000 > > > python vocoder.py -i ./mels-pre/ -o...