hexgrad

/

Kokoro-82M

like 210

I didn't design the Space to be duplicated because at the time Kokoro was not yet open weights, however the inference code and UI for v0.19 should be there (& MIT) so if you are fine with wrangling a bit code around, you should be able to have your own WebUI or cloned Space.

If you're not comfortable with editing code, you'd need to wait for someone else to do it, as I'm not sure when I'll be able to get to that.

There are also 2 projects I found about yesterday, which might help you out:

• https://github.com/remsky/Kokoro-FastAPI
• https://github.com/mateogon/pdf-narrator

Thanks Buddy Its Helped Alot

Hey @liaqateagle ,just as an FYI, a simple gradio webui utility has been added on https://github.com/remsky/Kokoro-FastAPI

got this error when running with docker :

fatal: not a git repository (or any parent up to mount point /app)
model-fetcher-1 | Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set).
kokoro-tts-1 | Traceback (most recent call last):
kokoro-tts-1 | File "/usr/local/bin/uvicorn", line 8, in
kokoro-tts-1 | sys.exit(main())
kokoro-tts-1 | File "/usr/local/lib/python3.10/dist-packages/click/core.py", line 1161, in call
kokoro-tts-1 | return self.main(*args, **kwargs)
kokoro-tts-1 | File "/usr/local/lib/python3.10/dist-packages/click/core.py", line 1082, in main
kokoro-tts-1 | rv = self.invoke(ctx)
kokoro-tts-1 | File "/usr/local/lib/python3.10/dist-packages/click/core.py", line 1443, in invoke
kokoro-tts-1 | return ctx.invoke(self.callback, **ctx.params)
kokoro-tts-1 | File "/usr/local/lib/python3.10/dist-packages/click/core.py", line 788, in invoke
kokoro-tts-1 | return __callback(*args, **kwargs)
kokoro-tts-1 | File "/usr/local/lib/python3.10/dist-packages/uvicorn/main.py", line 412, in main
kokoro-tts-1 | run(
kokoro-tts-1 | File "/usr/local/lib/python3.10/dist-packages/uvicorn/main.py", line 579, in run
kokoro-tts-1 | server.run()
kokoro-tts-1 | File "/usr/local/lib/python3.10/dist-packages/uvicorn/server.py", line 66, in run
kokoro-tts-1 | return asyncio.run(self.serve(sockets=sockets))
kokoro-tts-1 | File "/usr/lib/python3.10/asyncio/runners.py", line 44, in run
kokoro-tts-1 | return loop.run_until_complete(main)
kokoro-tts-1 | File "/usr/lib/python3.10/asyncio/base_events.py", line 649, in run_until_complete
kokoro-tts-1 | return future.result()
kokoro-tts-1 | File "/usr/local/lib/python3.10/dist-packages/uvicorn/server.py", line 70, in serve
kokoro-tts-1 | await self._serve(sockets)
kokoro-tts-1 | File "/usr/local/lib/python3.10/dist-packages/uvicorn/server.py", line 77, in _serve
kokoro-tts-1 | config.load()
kokoro-tts-1 | File "/usr/local/lib/python3.10/dist-packages/uvicorn/config.py", line 435, in load
kokoro-tts-1 | self.loaded_app = import_from_string(self.app)
kokoro-tts-1 | File "/usr/local/lib/python3.10/dist-packages/uvicorn/importer.py", line 22, in import_from_string
kokoro-tts-1 | raise exc from None
kokoro-tts-1 | File "/usr/local/lib/python3.10/dist-packages/uvicorn/importer.py", line 19, in import_from_string
kokoro-tts-1 | module = importlib.import_module(module_str)
kokoro-tts-1 | File "/usr/lib/python3.10/importlib/init.py", line 126, in import_module
kokoro-tts-1 | return _bootstrap._gcd_import(name[level:], package, level)
kokoro-tts-1 | File "", line 1050, in _gcd_import
kokoro-tts-1 | File "", line 1027, in _find_and_load
kokoro-tts-1 | File "", line 1006, in _find_and_load_unlocked
kokoro-tts-1 | File "", line 688, in _load_unlocked
kokoro-tts-1 | File "", line 883, in exec_module
kokoro-tts-1 | File "", line 241, in _call_with_frames_removed
kokoro-tts-1 | File "/app/api/src/main.py", line 13, in
kokoro-tts-1 | from .services.tts import TTSModel, TTSService
kokoro-tts-1 | File "/app/api/src/services/init.py", line 1, in
kokoro-tts-1 | from .tts import TTSModel, TTSService
kokoro-tts-1 | File "/app/api/src/services/tts.py", line 12, in
kokoro-tts-1 | from kokoro import generate, tokenize, phonemize, normalize_text
kokoro-tts-1 | ModuleNotFoundError: No module named 'kokoro'

Hey @liaqateagle

Can you see whether the huggingface pull was successful?

e.g. were the following directory and files created?

• ..\Kokoro-FastAPI\Kokoro-82M\
  --- kokoro.py
  --- kokoro-v0_19.pth
  --- etc

I'll take a look if I have any bugs with that part of the compose, maybe will switch it to a manual operation as it's been a little touchy
But if those files aren't there, try running the git clone https://huggingface.co/hexgrad/Kokoro-82M command in the same location as the docker-compose first.

hi buddy i have fixed it it was creating Kokoro-82M\Kokoro-82M
i moved files from Kokoro-82M\Kokoro-82M\ to Kokoro-82M\