Hugging Face's logo

Spaces:
WeReCooking
/
ACE-Step-CPU
Running

App Files Community
ACE-Step-CPU / Dockerfile

Commit History

  1. add fast captioning module (CLAP + faster-whisper + Silero VAD), update deps

    Nekochu
    4619f39

  1. remove XL checkpoint download (OOMKilled build, training uses standard turbo)

    Nekochu
    6d9fb39

  2. fix: save_every_n_epochs=0, add demucs-infer to Dockerfile, debug adapter dir

    Nekochu
    0e27e49

  1. major update: PyTorch inference, Gradio 6, session isolation, /understand captioning⁠

    Nekochu
    ff239f5

  1. add auto-captioning (BPM/key/signature via librosa), add librosa+mutagen deps

    Nekochu
    1d42836

  2. switch training to standard turbo (11s/epoch), auto-select standard GGUF for LoRA inference

    Nekochu
    c0f2a13

  3. fix: train on XL turbo (matches XL GGUF for inference), add XL checkpoint download

    Nekochu
    372f08e

  4. copy train_engine.py into Docker image

    Nekochu
    5fe3c53

  5. Side-Step training engine, tested locally on CPU⁠

    Nekochu
    a07b39d

  6. switch back to 1.7B LM (fastest at 269s, 0.6B was 936s)

    Nekochu
    5e95353

  7. swap LM to 0.6B Q8_0 for speed test

    Nekochu
    88b9223

  1. swap LM 4B->1.7B Q8_0 for faster CPU inference

    Nekochu
    b14d3e8

  2. remove accelerate (causes meta tensors), clean up patches

    Nekochu
    153f929

  1. add einops + vector_quantize_pytorch for model loading

    Nekochu
    88ca206

  2. pre-download training checkpoints at build time

    Nekochu
    c37e80e

  3. pin diffusers==0.30.3 for torch 2.4.x compat

    Nekochu
    f59d542

  4. pin torchaudio==2.4.0 (before torchcodec default backend)

    Nekochu
    148cd6b

  5. force uninstall torchcodec (torchaudio dependency, broken on Ubuntu 22.04)

    Nekochu
    9822bee

  6. remove torchcodec (needs FFmpeg 5.x), keep soundfile backend

    Nekochu
    63d7cb8

  7. add torchcodec + ffmpeg for torchaudio audio loading

    Nekochu
    f1f383b

  8. add soundfile + libsndfile1 for torchaudio backend

    Nekochu
    dd7a793

  1. add all training deps: diffusers lightning numpy tensorboard

    Nekochu
    eb1b926

  2. add torchaudio dep for training preprocess

    Nekochu
    9dc4031

  3. add loguru dep for training preprocess

    Nekochu
    d89a224

  4. install gradio[mcp] for mcp_server support

    Nekochu
    eb0e7f8

  5. add git to runtime image for ace-step source clone

    Nekochu
    0f0b7b0

  6. add LoRA training, fix css kwarg

    Nekochu
    625132a

  7. copy all shared libs including versioned symlinks

    Nekochu
    567a93e

  8. add pkg-config for BLAS detection

    Nekochu
    fd22b19

  9. switch to acestep.cpp GGUF: XL turbo Q4_K_M + 4B LM Q5_K_M

    Nekochu
    fccaf48