v1.1.8

What's Changed

• Add request queue to limit concurrent generation requests by @BBuf in #535
• News: 🔥🔥SGLang Diffusion x Cache-DiT ready!🔥🔥 by @DefTruth in #536
• feat: optimize async fp8 ulysses attn by @DefTruth in #537
• chore: delete useless code in serving by @BBuf in #539
• feat: make all_to_all comm unified by @DefTruth in #538
• feat: add refresh cache context api by @DefTruth in #542
• feat: support image edit model in serving by @BBuf in #541
• chore: simplify ulysses async flag by @DefTruth in #543
• chore: re-registered sage attn backend by @DefTruth in #544
• feat: support any head num for ulysses by @DefTruth in #546
• feat: support uneven heads in ulysses w/o padding by @triple-Mu in #547
• chore: refactor FLUX.2 image editing tests by @BBuf in #548
• feat: Add ulysses for any heads w/o padding by @DefTruth in #549
• feat: add envs manager for cache-dit by @DefTruth in #550
• fix ulysses fp8 and uneven head conflicts by @DefTruth in #552

Full Changelog: v1.1.7...v1.1.8