From 9116a55322d95631dc5cc4b738984603dca52f16 Mon Sep 17 00:00:00 2001 From: Vadim Gimpelson Date: Thu, 26 Mar 2026 02:33:49 +0400 Subject: [PATCH] adjust expected resultand tol Signed-off-by: Vadim Gimpelson --- tests/evals/gsm8k/configs/Qwen3.5-35B-A3B-DEP2.yaml | 4 ++-- tests/evals/gsm8k/configs/Qwen3.5-35B-A3B-FP8-DEP2.yaml | 4 ++-- tests/evals/gsm8k/configs/Qwen3.5-397B-A17B-NVFP4-DEP2.yaml | 4 ++-- 3 files changed, 6 insertions(+), 6 deletions(-) diff --git a/tests/evals/gsm8k/configs/Qwen3.5-35B-A3B-DEP2.yaml b/tests/evals/gsm8k/configs/Qwen3.5-35B-A3B-DEP2.yaml index 79b85b1e3dc..55a134ad9bd 100644 --- a/tests/evals/gsm8k/configs/Qwen3.5-35B-A3B-DEP2.yaml +++ b/tests/evals/gsm8k/configs/Qwen3.5-35B-A3B-DEP2.yaml @@ -1,6 +1,6 @@ model_name: "Qwen/Qwen3.5-35B-A3B" -accuracy_threshold: 0.85 -tolerance: 0.02 +accuracy_threshold: 0.84 +tolerance: 0.03 num_questions: 1319 num_fewshot: 5 server_args: >- diff --git a/tests/evals/gsm8k/configs/Qwen3.5-35B-A3B-FP8-DEP2.yaml b/tests/evals/gsm8k/configs/Qwen3.5-35B-A3B-FP8-DEP2.yaml index 52388ec872f..7a36052e35b 100644 --- a/tests/evals/gsm8k/configs/Qwen3.5-35B-A3B-FP8-DEP2.yaml +++ b/tests/evals/gsm8k/configs/Qwen3.5-35B-A3B-FP8-DEP2.yaml @@ -1,6 +1,6 @@ model_name: "Qwen/Qwen3.5-35B-A3B-FP8" -accuracy_threshold: 0.81 -tolerance: 0.02 +accuracy_threshold: 0.79 +tolerance: 0.03 num_questions: 1319 num_fewshot: 5 server_args: >- diff --git a/tests/evals/gsm8k/configs/Qwen3.5-397B-A17B-NVFP4-DEP2.yaml b/tests/evals/gsm8k/configs/Qwen3.5-397B-A17B-NVFP4-DEP2.yaml index 6bf0105e794..cd35790c345 100644 --- a/tests/evals/gsm8k/configs/Qwen3.5-397B-A17B-NVFP4-DEP2.yaml +++ b/tests/evals/gsm8k/configs/Qwen3.5-397B-A17B-NVFP4-DEP2.yaml @@ -1,6 +1,6 @@ model_name: "nvidia/Qwen3.5-397B-A17B-NVFP4" -accuracy_threshold: 0.89 -tolerance: 0.02 +accuracy_threshold: 0.88 +tolerance: 0.03 num_questions: 1319 num_fewshot: 5 server_args: >-