diff --git a/recipes/glm-5-3-flash-4x170hx-tp4-pp4-morrowmake-141/recipe.yaml b/recipes/glm-5-3-flash-4x170hx-tp4-pp4-morrowmake-141/recipe.yaml index d1683fff1..62cf20c71 100644 --- a/recipes/glm-5-3-flash-4x170hx-tp4-pp4-morrowmake-141/recipe.yaml +++ b/recipes/glm-5-3-flash-4x170hx-tp4-pp4-morrowmake-141/recipe.yaml @@ -44,8 +44,6 @@ config: - '--port' - '8000' - '--pipeline-parallel-size' - - '1' - - '--pipeline-parallel-size' - '4' - '--max-model-len' - '262144' @@ -63,6 +61,8 @@ config: - glm47 - '--speculative-config' - '{"method":"dflash","model":"/models/GLM-5.3-Flash-DFlash2","num_speculative_tokens":3}' + - '--block-size' + - '4608' - '--prefill-chunk-with-decodes' - '384' - '--max-num-partial-prefills'