EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[BugFix] fix qwen3-next compilation error (#7977)

Browse Source 
### What this PR does / why we need it?
fix qwen3-next compilation error

- vLLM version: v0.18.0
- vLLM release0.18.0:
445dc7196f
---------
Signed-off-by: cvSoldier <610496306@qq.com>
This commit is contained in:
cvSoldier
2026-04-03 20:03:39 +08:00
committed by GitHub
parent 81c6f51a45
commit 6c19270498
2 changed files with 59 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
vllm_ascend/ops/triton/mamba/causal_conv1d.py
View File

@@ -164,7 +164,6 @@ def extract_last_width(x, start_loc, width):
"stride_x_seq",
"stride_x_token",
"stride_conv_state_seq",
"stride_conv_state_tok",
"stride_state_indices",
"stride_o_seq",
"stride_o_token",
@@ -195,7 +194,7 @@ def _causal_conv1d_update_kernel_npu_tiled(
stride_w_width: tl.constexpr,
stride_conv_state_seq,
stride_conv_state_dim: tl.constexpr,
stride_conv_state_tok,
stride_conv_state_tok: tl.constexpr,
stride_state_indices,
stride_o_seq,
stride_o_dim: tl.constexpr,