EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

[BugFix][Qwen3-VL]: fix cu_seqlens in qwen3-vl (#11458)

Browse Source
This commit is contained in:
Zheng Wengang
2025-10-15 13:16:49 +08:00
committed by GitHub
parent 32803fb279
commit b2c8566920
1 changed files with 5 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
python/sglang/srt/models/qwen3_vl.py
View File

@@ -452,13 +452,15 @@ class Qwen3_VisionTransformer(nn.Module):
position_embeddings = (emb.cos(), emb.sin()) position_embeddings = (emb.cos(), emb.sin())
# compute cu_seqlens # compute cu_seqlens
cu_seqlens = torch.repeat_interleave(
grid_thw[:, 1] * grid_thw[:, 2], grid_thw[:, 0]
).cumsum(dim=0)
cu_seqlens = torch.cat( cu_seqlens = torch.cat(
[ [
torch.tensor([0], device=grid_thw.device), torch.zeros(1, dtype=torch.int32, device=cu_seqlens.device),
(grid_thw[:, 0] * grid_thw[:, 1] * grid_thw[:, 2]).cumsum(dim=0), cu_seqlens.to(torch.int32),
] ]
) )
cu_seqlens = torch.cat([cu_seqlens.new_zeros(1), cu_seqlens])
# max_seqlen, seqlens = self.compute_attn_mask_seqlen(cu_seqlens) # max_seqlen, seqlens = self.compute_attn_mask_seqlen(cu_seqlens)
x = x.unsqueeze(1) x = x.unsqueeze(1)