Commit History

  1. Add /normal16 endpoint: raw float16 normals (.npy) to avoid 8-bit quantisation grain

    Verified signature
    c6a2f5f

  1. Return normal map as lossless PNG via gr.File (avoid webp orange-peel)

    Verified signature
    5836176
  2. Return depth as 16-bit PNG via gr.File (preserve full precision)

    Verified signature
    5b41476
  3. Fix 16-bit->8-bit depth conversion

    Verified signature
    cd54749
  4. Switch to ZeroGPU hardware (model requires float16 = GPU)

    Verified signature
    f76e55a
  5. Add show_error=True to see inference error details

    Verified signature
    3137796
  6. Fix bool-schema bug in gradio_client (crash on /gradio_api/info)

    Verified signature
    f368279
  7. Bind to 0.0.0.0:7860 so HF proxy can reach the server

    Verified signature
    8e61a1b
  8. Pin transformers==4.44.0 (has FLAX_WEIGHTS_NAME); add torchvision

    Verified signature
    e44a10a
  9. Add cached_download + HfFolder shims for hub>=0.30 compat

    Verified signature
    35745a5
  10. Add HfFolder compat shim for huggingface_hub>=0.30

    Verified signature
    5c3d6e6
  11. Fix ResolutionImpossible: drop huggingface_hub pin + torch pin (HF base provides)

    Verified signature
    59e92eb
  12. No gradio-client pin - let gradio 5.x sdk injection handle it

    Verified signature
    61798c0
  13. Remove gradio-client pin (conflicts with gradio 5.x sdk injection)

    Verified signature
    0c1f775
  14. Bump gradio sdk to 5.6.0 (python 3.10 already pinned, audioop safe)

    Verified signature
    27ea63b
  15. Override bundled gradio-client to fix json-schema bool bug

    Verified signature
    43b37e8
  16. Add torchvision; loose torch pin

    Verified signature
    b320327
  17. Remove unused import tensorboard

    Verified signature
    4a6aec2
  18. Pin python_version to 3.10 (audioop removed in 3.13)

    Verified signature
    207c75a
  19. Fix syntax errors from escape mangling

    Verified signature
    01bc874
  20. Slim requirements: drop torch (preinstalled), drop gradio (sdk handles)

    Verified signature
    6ea6120
  21. Revert sdk_version to 4.44.0

    Verified signature
    933ba56
  22. Match haodongli build deps (torch 2.5.1, gradio 4.44)

    Verified signature
    5c08523
  23. Bump sdk_version to gradio 5.6.0

    Verified signature
    54ccc56
  24. Slim requirements; upgrade gradio to v5

    Verified signature
    f44fb38
  25. Minimal dual-endpoint app: /depth + /normal

    Verified signature
    738c809

  1. saving as .tiff didn't work - it doesnt seem to calculate the depth

    Verified signature
    ec99a99
  2. it's .TIFF, not TIFF

    Verified signature
    d5dcbc5
  3. Trying to force the image to be saved as TIFF

    Verified signature
    e28a3c5
  4. Revert num_inference_steps=1 from 3. Didnt seem to change much

    Verified signature
    15ba24b
  5. num_inference_steps=3, from 1 - trying to remove the noise in the generated image

    Verified signature
    83b9f2e
  6. Update to export depth maps for depth estimation as 16bits, not 8 bits

    Verified signature
    824ae52

  1. Update utils/image_utils.py

    Verified signature
    e94a31b

  1. move back

    79ca2bd

  1. update video depth

    b6df43e

  1. huggingface_hub==0.25.0

    572240d
  2. update lotus-depth-g-v2-1-disparity

    b5ad4c4

  1. add resizing

    f562867
  2. add resizing

    fe3664d
  3. add resizing

    5f0675a

  1. update for video depth

    a47128f
  2. update gitattributes

    4943e69
  3. delete old video

    d1b92d2

  1. update github

    e2ac6fb
  2. update X

    02075c2

  1. update port

    f8a8adc
  2. remove vid

    8020585

  1. update for zerogpu

    6fcc783
  2. update for zerogpu

    25fc9d6
  3. add transformers.utils.move_cache()

    ea0a1d3