Vimos Tan

Can you illustrate more on your task and what you have tried?

I added the file to the shared link, https://drive.google.com/drive/folders/1r2G1KNmITPNsm1jZK0XWtf6jXI7MaZfo?usp=sharing

都是生成的tokenize之后的ids文件，通过preprocess生成的，如果不需要这一步，可以跳过去直接tokenize之后传递给模型

I will upload to this link, https://drive.google.com/drive/folders/1r2G1KNmITPNsm1jZK0XWtf6jXI7MaZfo?usp=sharing

加了这个也不行，看来超了容许的量了 ``` git lfs install git clone https://github.com/wac81/vits_chinese.git ``` ``` Error downloading object: logs/woman_csmsc/G_100000.pth (9c70b3c): Smudge error: Error downloading logs/woman_csmsc/G_100000.pth (9c70b3cfea4e35c3aecc3ba854848ea14b974e67417efbadf60e365ebaca91c7): batch response: This repository is over its data quota....

I also want to ask if there are simple APIs for testing enctryption. A similar function in `zipfile`: ``` with zipfile.ZipFile(archive_stream) as rd: for rf in rd.infolist(): is_encrypted = rf.flag_bits...

标点符号宽度和word不同，影响了每行的字数。

我也发现用gpu为零的情况，换cpu就没有问题。 看来这里至少需要个提示。

The vocab is constructed from these files, please refer to https://github.com/VisualJoyce/ChengyuBERT/blob/88dbc1cd4c6904133b3363248f2bdc5cc10a9099/chengyubert/data/__init__.py#L116C1-L144C25

Thank you! I will notice them about the issue.