Spaces:
Running
Running
Commit History
-
Auto-detect available ffmpeg codec (libx264/libopenh264/mpeg4 fallback) Running
Verified signature4c2f9e8 -
Remove -crf and complex vf filter - use simple -pix_fmt yuv420p
Verified signature2918b43 -
Fix ffmpeg: use shell=True, reorder -crf option
Verified signature0e93a6d -
Fix ffmpeg: use subprocess.run with error checking, absolute temp path
Verified signature471ac47 -
Add debug listing to download endpoint
Verified signature1cfc19b -
Add custom /api/download endpoint to bypass Gradio file serving 403
Verified signature4dc9390 -
Add debug logging for file access
Verified signature855ad26 -
Debug: allowed_paths=['/'] + logging
Verified signature8716ff6 -
Add allowed_paths /tmp for file serving
Verified signatureafa7564 -
Set GRADIO_TEMP_DIR=/tmp for file serving
Verified signature85c4867 -
Save results to Gradio cache dir for proper file serving
Verified signaturef8501f2 -
Save results to temp dir for Gradio 5 file serving
Verified signaturee23be24 -
Fix allowed_paths with absolute path for results directory
Verified signature274d5f7 -
Add allowed_paths for results directory
Verified signatureFix file download 403f3bae02 -
fix: use direct paste-back instead of get_image (requires face parsing model we don't init)
Verified signature6286194 -
fix: use proper tensor dtype conversion in inference loop (FloatTensor fails on CUDA half tensors)
Verified signature4054d5d -
Fix gradio_client schema bug + update app
Verified signature1dd102d -
fix: correct get_whisper_chunk parameter order (device was getting librosa_length)
Verified signatureb1a0564 -
fix: update model loading API
Verified signaturec693a02 -
fix: revert to eager model loading — L4 GPU loads models fast, deferred loading caused inference errors
Verified signaturee520d04 -
fix: COPY app.py into Docker image (was using GitHub clone, not our custom version)
Verified signaturefa2112e
-
fix: defer model loading to first inference — unblocks Gradio startup
Verified signaturee8ed3bd -
fix: simplify entrypoint — models pre-downloaded in Docker image
Verified signaturec3b200d -
fix: download models at Docker build time to avoid 30-min startup timeout
Verified signature00e1a5a -
fix: re-add SyncNet download (required by MuseTalk)
Verified signatureab771f9 -
fix: remove debug=True from launch
Verified signature06a770e -
fix: parallelize model downloads to fit 30-min startup window
Verified signature27ee4a3 -
fix: add download retry, validate models, use exec for clean process mgmt
Verified signaturea6039bd -
fix: resolve NameError, align model paths to models/, remove @spaces.GPU
Verified signature03c94e3 -
fix: download diffusion_pytorch_model.safetensors for SD VAE
Verified signature433b0f3 -
fix: pin gdown<5.0 to fix missing gdown.cli module
Verified signature43e7aa0 -
Fix: pin huggingface_hub<1.0 to fix CLI import error
Verified signature371c7a8 -
Fix: COPY --chown=user:user so chmod +x succeeds
Verified signature4e5152c -
Fix: chmod +x entrypoint.sh after COPY to fix permission denied
Verified signature02f26b6 -
Fix Dockerfile: COPY entrypoint.sh into container for model downloads
Verified signaturea8cbb0d -
Fix entrypoint: download model weights before starting app
Verified signature953b87f -
Fix: download model weights at startup
Verified signaturecc26e9f
-
Update packages.txt
anchorVerified signature59976c1 -
Update requirements.txt
anchorVerified signature4737142 -
Update requirements.txt
anchorVerified signatureaf39918 -
Update requirements.txt
anchorVerified signature36eae46
-
Update requirements.txt
Verified signaturef8bf34c -
Update Dockerfile
anchorVerified signature1a80aee -
Upload install_ffmpeg.sh
anchorVerified signatureb68506c -
Update packages.txt
Verified signature4dab231 -
Update packages.txt
anchorVerified signatureaca9660 -
Update packages.txt
anchorVerified signature595ac18 -
Update packages.txt
anchorVerified signaturead766a6 -
Update packages.txt
anchorVerified signatureb4d16f5 -
Update requirements.txt
anchorVerified signature1370535
