[chore] add docs for macos

[Tps-F]

Is this PR adding new feature or fix a BUG?

Fix BUG.

Is this pull request related to any issue? If yes, please link the issue.

https://discord.com/channels/1214047546020728892/1214047661758484531/1284118763767070750
#461 (comment)

I just added the installation instructions on macOS to the documentation.

[leng-yue]

Any update so we can merge this?

[Tps-F]

Sorryfor late, I have been busy studying for college exams lately and have not yet been able to look into the cause of the slowdown on the mps device.

I will add some warnings tomorrow and then I hope you will check this PR as is!

I will create another PR later on about faster inference speeds.

[Tps-F]

I'm really Sorry for the delay.
I have added about usage and current warnings.
For Chinese and Portuguese, I used a translator, so if there are any oddities, I would appreciate it if you could correct them!

[leng-yue]

LGTM, thanks for your contribution :)

[Tps-F]

Of course, thank you!
I want to keep contributing to this project from now on as well

[brainstorm]

If you want to perform inference on MPS, please add the --device mps flag.

Add the --device mps flag where exactly (which command)? Please provide an example.

EDIT:

Ok, I guess that flag goes to the command found in entry entrypoint.sh, so:

python tools/webui.py --device mps

Also (should I open a separate issue?): cachetools should be installed via pyproject.toml's runtime dependencies:

% python tools/webui.py --device mps
Traceback (most recent call last):
  File "/Users/rvalls/dev/personal/fish-speech/tools/webui.py", line 25, in <module>
    from tools.api import decode_vq_tokens, encode_reference
  File "/Users/rvalls/dev/personal/fish-speech/tools/api.py", line 42, in <module>
    from cachetools import LRUCache, cached
ModuleNotFoundError: No module named 'cachetools'