Alexander Korolev 1d0bb43ece mimo26: vLLM B12X A/B lane recipe + sglang 4x fp8-KV trial rollback 3 هفته پیش
..
reports 4ace7f4d06 bench roce-quality: 168/184 (91/100), resp 79, decode +10-15% — RoCEnante is quality-neutral 3 هفته پیش
recipe.toml 1d0bb43ece mimo26: vLLM B12X A/B lane recipe + sglang 4x fp8-KV trial rollback 3 هفته پیش