Commit History

  1. Fix FileNotFoundError: Use shutil.copy instead of move so Gradio doesn't lose the temp image on second run

    807350d
  2. Fix Gradio FFmpeg error: remove format parameter to prevent in-place re-encoding

    b445b47
  3. Translate Gradio settings to Spanish and make them clearer

    4252285
  4. Fix FileNotFoundError: create workspace directory if it doesn't exist

    c2ab61f
  5. Fix TypeError: Convert predictions to CPU numpy before np.savez

    617e6a3
  6. Fix numpy 2.0 ValueError: inhomogeneous array shape in trans_params

    e6da31e
  7. Add numpy monkey patch for facexlib compatibility on modern Python/NumPy 2.x

    321b228
  8. Fix numpy 2.0 AttributeError: remove obsolete VisibleDeprecationWarning

    c424d4c
  9. Fix numpy build error in Colab by relaxing version pinning

    0328694
  10. Execute Ponytail audit cuts and inject autocast FP16 for extreme GPU acceleration

    5f77949
  11. Refactor for GPU: Remove CPU monkey patches, add CuDNN benchmark, add CuPy, enable Gradio share

    0fe0e16
  12. Disable torch.compile completely to fix spectral_norm crashes on CPU

    ad68d7c
  13. Suppress torch._dynamo errors on CPU to fix spectral_norm compile failure

    2c5882f
  14. Fix facexlib hardcoded CUDA device and optimize Dockerfile caching

    7578d72
  15. Fix AnimateFromCoeff missing bf16 argument

    81003d9
  16. Monkey patch torch.load for CPU-only inference in Hugging Face Spaces

    97a2f41
  17. Fix Gradio 4 Tabs rendering crash

    7db19ec
  18. Bind Gradio to 0.0.0.0 to fix Docker Space starting timeout

    8770772
  19. Fix Gradio 4 compatibility and dependency conflicts

    f6e4156
  20. Fix UI hang by pinning Gradio to 3.50.2 and removing analytics_enabled

    784aeaa
  21. Download models during docker build to fix missing checkpoints

    8cd20e7
  22. Disable TTS to avoid missing dependency

    3c05cdd
  23. Fix gradio 4.0 deprecation errors in app.py

    eb5c2bc
  24. Fix scipy version for python 3.10

    a8da5cd
  25. Fix Hugging Face metadata in README

    35500eb
  26. Migrate xtalker to Dockerized Hugging Face Space with GPU support and LFS

    31bd91f

  1. Update code

    63b79e9
  2. Update code

    d200feb
  3. Update code

    8686eb7
  4. Update code

    9a5ca1f
  5. Use precompiled llama-cpp-python wheel

    36be99c

  1. Add Hugging Face configuration in README.md

    ef50184
  2. Add application file

    8b20d23