From 4bf06be985c64edbb0c8b0f994ceff787ce84b09 Mon Sep 17 00:00:00 2001 From: "Kevin H. Luu" Date: Tue, 1 Sep 2026 13:24:46 -0700 Subject: [PATCH] [CI] Disable CUDA graphs for GLM PCP evals (#54745) Signed-off-by: Kevin Luu <51931015+khluu@users.noreply.github.com> --- tests/evals/gsm8k/configs/GLM-5.2-NVFP4-TP1-PCP4-EP.yaml | 2 +- tests/evals/gsm8k/configs/GLM-5.2-NVFP4-TP2-PCP2-EP.yaml | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/tests/evals/gsm8k/configs/GLM-5.2-NVFP4-TP1-PCP4-EP.yaml b/tests/evals/gsm8k/configs/GLM-5.2-NVFP4-TP1-PCP4-EP.yaml index 9679fbebb7..aabe783f65 100644 --- a/tests/evals/gsm8k/configs/GLM-5.2-NVFP4-TP1-PCP4-EP.yaml +++ b/tests/evals/gsm8k/configs/GLM-5.2-NVFP4-TP1-PCP4-EP.yaml @@ -4,7 +4,7 @@ num_questions: 1319 num_fewshot: 5 max_concurrency: 100 server_args: >- - --compilation-config '{"cudagraph_mode":"PIECEWISE"}' + --compilation-config '{"cudagraph_mode":"NONE"}' --max-model-len 4096 --max-num-batched-tokens 32768 --safetensors-load-strategy prefetch diff --git a/tests/evals/gsm8k/configs/GLM-5.2-NVFP4-TP2-PCP2-EP.yaml b/tests/evals/gsm8k/configs/GLM-5.2-NVFP4-TP2-PCP2-EP.yaml index 040e076edb..72f7fdf1ae 100644 --- a/tests/evals/gsm8k/configs/GLM-5.2-NVFP4-TP2-PCP2-EP.yaml +++ b/tests/evals/gsm8k/configs/GLM-5.2-NVFP4-TP2-PCP2-EP.yaml @@ -4,7 +4,7 @@ num_questions: 1319 num_fewshot: 5 max_concurrency: 100 server_args: >- - --compilation-config '{"cudagraph_mode":"PIECEWISE"}' + --compilation-config '{"cudagraph_mode":"NONE"}' --max-model-len 4096 --max-num-batched-tokens 32768 --safetensors-load-strategy prefetch