[BUG REPORT] Lock gradio version

[melMass]

Describe the bug

While trying to get the server running again I reinstalled it but it won't run anymore.

To Reproduce Install gradio, either manually or using the install script this will install 4.10

Expected behavior Either lock the gradio req or fix the bug

Screenshots If applicable, add screenshots to help explain your problem.

Additional context

Windows 11

2023-12-16 14:01:05 | WARNING | xformers | A matching Triton is not available, some optimizations will not be enabled.
Error caught was: No module named 'triton'
2023-12-16 14:01:05 | WARNING | xformers | Triton is not available, some optimizations will not be enabled.
This is just a warning: No module named 'triton'
Traceback (most recent call last):
  File "audio-webui.git\main.py", line 59, in <module>
    from webui.modules.implementations.tts_monkeypatching import patch as patch1
  File "audio-webui.git\webui\modules\implementations\__init__.py", line 1, in <module>
    import webui.modules.implementations.ttsmodels as tts
  File "audio-webui.git\webui\modules\implementations\ttsmodels.py", line 15, in <module>
    from webui.modules.implementations.patches.bark_custom_voices import wav_to_semantics, generate_fine_from_wav, \
  File "audio-webui.git\webui\modules\implementations\patches\bark_custom_voices.py", line 8, in <module>
    from webui.modules.implementations.patches.bark_generation import generate_text_semantic_new, generate_coarse_new, generate_fine_new
  File "audio-webui.git\webui\modules\implementations\patches\bark_generation.py", line 7, in <module>
    from webui.ui.tabs import settings
  File "audio-webui.git\webui\ui\tabs\__init__.py", line 2, in <module>
    from .text_to_speech import text_to_speech
  File "audio-webui.git\webui\ui\tabs\text_to_speech.py", line 2, in <module>
    from gradio.components import IOComponent
ImportError: cannot import name 'IOComponent' from 'gradio.components' (.venv\lib\site-packages\gradio\components\__init__.py)
cannot import name 'IOComponent' from 'gradio.components' (.venv\lib\site-packages\gradio\components\__init__.py)
Your install might have failed to install one of the requirements, are you missing a package?
You can read common issues at https://github.com/gitmylo/audio-webui/wiki/common-issues

[gitmylo]

Do you know at which version it stopped working? If I had to guess it would be with the version 4 release, since that's around where the github removed IOComponent from the gradio.components import.

Although audio-webui locks gradio at 3.35.2 the last version that could still work is 3.50.2.

If you use the built in automatic installer, it will install 3.35.2.

[melMass]

Do you know at which version it stopped working?

Unfortunately I don't

Although audio-webui locks gradio at 3.35.2 the last version that could still work is 3.50.2.

Thanks let me try now and report

If you use the built in automatic installer, it will install 3.35.2.

I avoid automatic stuff since I like to manage my envs but I did follow the paths to makes sure to follow these, may I ask where this is enforced I don't see it?

EDIT: found it here

[gitmylo]

The requirement info is stored here: setup_tools/magicinstaller/requirements/init.py

[melMass]

It works! I also had to change the port used by torch distributed somehow (it failed to bind otherwise) And trying the torch-pepe based extraction crashed the server with a symbol not found:

gin_channels: 256 self.spk_embed_dim: 109
<All keys matched successfully>
LLVM ERROR: Symbol not found: __svml_cosf8_ha