FlashPortrait Current Facial SOTA model building

https://github.com/Francis-Rings/FlashPortrait

FlashPortrait라고 표정 분야의 SOTA 모델이 오픈소스로 출시되었다고 해서

이를 빌드하고 사용해보려고 한다.

conda create -n FlashPortrait -y python==3.12.12
conda activate FlashPortrait
pip install torch==2.6.0 torchvision==0.21.0 torchaudio==2.6.0 --index-url https://download.pytorch.org/whl/cu124
git clone -b main https://github.com/Francis-Rings/FlashPortrait.git
cd FlashPortrait
pip install -r requirements.txt
pip install flash-attn==2.7.4.post1 --no-build-isolation
pip install "huggingface_hub[cli]"
mkdir checkpoints
huggingface-cli download FrancisRing/FlashPortrait --local-dir ./checkpoints/FlashPortrait
huggingface-cli download Wan-AI/Wan2.1-I2V-14B-720P --local-dir ./checkpoints/Wan2.1-I2V-14B-720P
pip install scipy
pip install loguru
pip install xfuser

sed -i 's|/path/FlashPortrait|/root/FlashPortrait|g' fast_infer.py
python fast_infer.py \
  --GPU_memory_mode sequential_cpu_offload \
  --ckpt_dir /root/FlashPortrait/checkpoints/Wan2.1-I2V-14B-720P

---

실행은 되는데 시스템 메모리가 스왑해도 모자른지 중간에 WSL이 튕긴다…

나중에 메모리가 되면 시도하겠음