Sparge Attention (cogvideox-2b, want2v-1.3B)

GitHub - thu-ml/SpargeAttn: SpargeAttention: A training-free sparse attention that can accelerate any model inference.

GitHub - shenduldh/ComfyUI-Lightning: Accelerate inference in Flux and Sana for ComfyUI.

Xiang-cd/sparge-attention-model-zoo В· Hugging Face