aiueo52/hyperburst-rtxpro6000

Qwen3.8-Flash-Next: fast single-stream decode on one RTX PRO 6000 Blackwell Max-Q — method, results, SGLang patches, benchmarks, and a guide to train your own MTP head

语言：Python · GitHub 星标：9 · Fork：1

在 GitHub 上查看原仓库