Commit History

Today

  1. fix(enhance): normalize FFA-Net input like the official test code⁠ Running

    IbProgrammmer +1: Claude Opus 5.5
    b123fa1
  2. fix(docker): upgrade pip and pin +cpu torch wheels so the torch install step succeeds⁠

    IbProgrammmer +1: Claude Opus 5.5
    0cb72a1
  3. feat(detect): use MMDetection RTMDet-s as the primary detector⁠

    IbProgrammmer +1: Claude Opus 5.5
    868be2c

Yesterday

  1. fix(enhance): load FFA-Net and Zero-DCE++ weights correctly⁠

    IbProgrammmer +1: Claude Opus 5.5
    80867d6

  1. feat(video): preserve audio track and fix video frame assembly⁠

    IbProgrammmer +1: Claude Sonnet 4.6
    e75f322
  2. feat(video): add /pipeline_video endpoint with frame-by-frame annotation and H264 output

    3c7c247
  3. Add HF Spaces README with Docker sdk configuration

    ea7968c

  1. fix(detect): raise confidence threshold 0.25→0.45 and iou 0.45 to cut false positives

    713c1ee

  1. fix: use /tmp/models for writable model cache, hardcode HF repo

    871bbe6
  2. fix: rename env vars to alphanumeric-only for HF Spaces secret naming rules

    bf6ac6d
  3. fix(startup): always load pretrained yolov8n fallback regardless of file existence

    9363b95
  4. fix(docker): replace libgl1-mesa-glx with libgl1 for Debian trixie

    911230f
  5. deploy: support all models (YOLO, FFA-Net, Zero-DCE++, Restormer, buffalo_l)⁠

    IbProgrammmer +1: Claude Sonnet 4.6
    a467a79
  6. deploy: add HF Hub model download on startup⁠

    IbProgrammmer +1: Claude Sonnet 4.6
    1e13785

  1. fix

    31931b6
  2. add deploy/ folder — HuggingFace Spaces inference API⁠

    IbProgrammmer +1: Claude Sonnet 4.6
    03c4042