Fetching metadata from the HF Docker repository...
-
Ensure /ready endpoint triggers safe model initialization
-
Fix ZeroGPU RuntimeError: eliminate import-time CUDA calls and add seamless CPU fallback
-
Fix Gradio Space UI: disable SSR mode, update SDK to 5.50.0, and rebuild Blocks layout
-
feat(app): invoke zero_client.startup_report() to notify ZeroGPU supervisor
-
fix(app): import spaces and register ZeroGPU marker before torch import
-
feat(app): add @spaces.GPU decorator for Hugging Face ZeroGPU runtime
-
fix(app): add port binding diagnostics and fallback
-
fix(server): add safe dlib face detection fallback if face_recognition raises SystemExit
-
fix(deps): add setuptools for pkg_resources used by face_recognition_models
-
fix(requirements): use resolve URL for dlib wheel in container build
-
fix(deps): track precompiled dlib Linux binary wheel via LFS to avoid build OOM
-
Deploy Gradio Deepfake Detection inference backend to Hugging Face Spaces
-
initial commit