类库
› FlashVSR
OpenImagingLab/FlashVSR
FlashVSR是一个基于扩散模型的实时视频超分辨率框架，通过单步扩散、局部约束稀疏注意力和小型条件解码器等创新，显著提升处理效率，支持对高分辨率视频进行流畅的超分辨率增强，适用于实时流媒体场景。
技术栈
查看全部依赖 (25)
依赖
NumPy
unknown
Pandas
unknown
accelerate
unknown
datasets
unknown
einops
unknown
ftfy
unknown
huggingface-hub
unknown
imageio
unknown
imageio-ffmpeg
unknown
matplotlib
unknown
opencv-python
unknown
opencv-python-headless
unknown
peft
unknown
pillow
unknown
protobuf
unknown
pytorch-lightning
unknown
safetensors
unknown
sentencepiece
unknown
torch
unknown
torchaudio
unknown
torchmetrics
unknown
torchsde
unknown
torchvision
unknown
tqdm
unknown
transformers
unknown
截图
