It’s a good idea! I’d also like to help when I have time. 
Tensorflow and Pytorch Chinese manuals are based on Markdown. It might be easier to write than RST, but we might want to keep the Chinese version align with the English version if we want to host it under Apache. I think we need to clarify whether this will be a part of TVM core manual development so it can be put in the Apache TVM repo, or a community’s effort to enhance TVM (in which case it can have its own github repo). @hogepodge, do you have suggestions and comments?
Btw, @BBuf has written a few great articles in Chinese about TVM with example code in this repo: https://github.com/BBuf/tvm_learn.