Hugging Face's logo

Spaces:
WeReCooking
/
ACE-Step-CPU
Running

App Files Community
ACE-Step-CPU / app.py

Commit History

  1. redirect training subprocess stderr to log file for debugging

    Nekochu
    e69e9ec

  2. run training as detached subprocess to survive Gradio session timeout

    Nekochu
    a4a86a8

  3. use bfloat16 precision for training to halve RAM usage

    Nekochu
    a0e1f4c

  4. stop ace-server during training to free RAM, restart after, add log visibility

    Nekochu
    c2cb0b9

  5. monkey-patch torchaudio.load to use soundfile backend

    Nekochu
    afccbc0

  1. add LoRA training, fix css kwarg

    Nekochu
    625132a

  2. full app: CLI + Gradio + training placeholder + MCP

    Nekochu
    72e4b69

  3. rewrite app.py for ace-server HTTP API, no torch

    Nekochu
    2dc2899

  4. clickable title

    Nekochu
    fe00878

  5. slim title

    Nekochu
    1e6f1c3

  6. compact grid UI, 4B default, show training model

    Nekochu
    4b376ab

  7. force float32 on training model for CPU

    Nekochu
    562fa54

  8. fix: LoRA rank param is 'r' not 'rank'

    Nekochu
    dece91f

  9. fix training: use LoRAConfigV2 instead of Union type alias

    Nekochu
    a6b0e73

  10. step 7: real app.py

    Nekochu
    dae27d4

  11. step 6: +numba (all deps)

    Nekochu
    2da29c2

  12. step 5: +vector-quantize-pytorch

    Nekochu
    3280257

  13. step 4: +peft +torchvision

    Nekochu
    27ddf91

  14. step 3: core deps without heavy ones

    Nekochu
    7de0ea2

  15. step label

    Nekochu
    f0eaee4

  16. step 1: bare gradio

    Nekochu
    5c82a90