From cf7923615c263480a24a7b45695801d5eea3e5d5 Mon Sep 17 00:00:00 2001 From: YC Yen-Ching Tseng Date: Thu, 6 Aug 2026 15:10:22 +0800 Subject: [PATCH] [AMD] Gate DFLASH non-greedy verify on the target-only kernel being registered (#33694) Co-authored-by: HaiShaw --- python/sglang/srt/speculative/dflash_utils.py | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/python/sglang/srt/speculative/dflash_utils.py b/python/sglang/srt/speculative/dflash_utils.py index 284ae5661..f5aa6e158 100644 --- a/python/sglang/srt/speculative/dflash_utils.py +++ b/python/sglang/srt/speculative/dflash_utils.py @@ -53,8 +53,7 @@ elif is_hip(): from sglang.kernels.ops.sampling.renorm_triton import ( top_p_renorm_probs_triton as top_p_renorm_prob, ) - - _DFLASH_SAMPLING_VERIFY_AVAILABLE = True + tree_speculative_sampling_target_only = None else: top_k_renorm_prob = None top_p_renorm_prob = None