DiTFastAttn: Attention Compression for Diffusion Transformer Models Paper • 2406.08552 • Published Jun 12, 2024 • 23