tests/evals/gsm8k/configs/Qwen3.5-35B-A3B-DEP2.yaml

model_name: "Qwen/Qwen3.5-35B-A3B"
accuracy_threshold: 0.84
tolerance: 0.03
num_questions: 1319
num_fewshot: 5
server_args: >-
  --max-model-len 4096
  --data-parallel-size 2
  --enable-expert-parallel
[Bugfix] Disable monolithic TRTLLM MoE for Renormalize routing (#37591) (#37605) Signed-off-by: Vadim Gimpelson <vadim.gimpelson@gmail.com> 2026-03-20 23:19:26 +04:00			`model_name: "Qwen/Qwen3.5-35B-A3B"`
[Bugfix] Fix DeepGemm E8M0 accuracy degradation for Qwen3.5 FP8 on Blackwell (#38083) Signed-off-by: Vadim Gimpelson <vadim.gimpelson@gmail.com> 2026-03-26 12:21:47 +04:00			`accuracy_threshold: 0.84`
			`tolerance: 0.03`
[Bugfix] Disable monolithic TRTLLM MoE for Renormalize routing (#37591) (#37605) Signed-off-by: Vadim Gimpelson <vadim.gimpelson@gmail.com> 2026-03-20 23:19:26 +04:00			`num_questions: 1319`
			`num_fewshot: 5`
			`server_args: >-`
			`--max-model-len 4096`
			`--data-parallel-size 2`
			`--enable-expert-parallel`