M$^2$-ViT: Accelerating Hybrid Vision Transformers with Two-Level Mixed Quantization

MQ
Re-assign community