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