A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 03:30 UTC
[Refactor] Remove useless pdb comment (#20266)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Wentao Ye committed
Jun 30, 2025 at 14:15 UTC
97d9524fe90ad5799cc11db4b4216fe3a30a07d6
1 file changed
-1
vllm/model_executor/layers/fused_moe/deep_gemm_moe.py
-1
@@ -141,7 +141,6 @@ class DeepGemmExperts(mk.FusedMoEPermuteExpertsUnpermute):
141
quant_out = _resize_cache(workspace13.view(dtype=torch.float8_e4m3fn),
142
(M_sum, N // 2))
143
mm2_out = _resize_cache(workspace2, (M_sum, K))
144
- # import pdb; pdb.set_trace()
144
145
dg.m_grouped_gemm_fp8_fp8_bf16_nt_contiguous(
146
(a1q, a1q_scale), (w1, w1_scale), mm1_out, expert_ids)