Run graph under torch.inference_mode() (fix inference-tensor inplace error)

Use tiled VAE decode to bound peak VRAM (fix NVML allocator assert)

Fix audio loading: use soundfile instead of torchaudio (torchcodec missing)

Upload requirements.txt with huggingface_hub

Upload requirements.txt with huggingface_hub

Delete __pycache__/app.cpython-311.pyc with huggingface_hub

Upload folder using huggingface_hub

Upload requirements.txt with huggingface_hub

Upload app.py with huggingface_hub

Upload requirements.txt with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload folder using huggingface_hub

initial commit