xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-09 04:54:56 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Bug] Fix Assertion error DeepEP/csrc/kernels/intranode.cu:928: 'false and Unsupported type' (#26532)

Browse Source 
Signed-off-by: yewentao256 <zhyanwentao@126.com>
This commit is contained in:
Wentao Ye 2025-10-13 18:26:37 -04:00 committed by GitHub
parent 577c72a227
commit 7200a21cd1
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 5 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
vllm/model_executor/layers/fused_moe/deepep_ht_prepare_finalize.py
View File

@ -336,7 +336,11 @@ class DeepEPHTPrepareAndFinalize(mk.FusedMoEPrepareAndFinalize):
apply_router_weight_on_input=apply_router_weight_on_input,
)
dbo_yield_and_switch_from_compute_to_comm()
assert fused_expert_output.dtype == torch.bfloat16, (
f"Expected fused_expert_output bfloat16, got {fused_expert_output.dtype}"
)
combined_x, _, event = self.buffer.combine(
# HT combine only supports BF16
x=fused_expert_output,
handle=handle,
topk_weights=None,

2
vllm/model_executor/layers/fused_moe/modular_kernel.py
View File

@ -984,7 +984,7 @@ class FusedMoEModularKernel(torch.nn.Module):
assert num_chunks == 0
workspace13 = None
workspace2 = None
fused_out = torch.empty_like(a1q)
fused_out = torch.empty_like(a1q, dtype=in_dtype)
else:
assert num_chunks > 0
workspace13, workspace2, fused_out = self._allocate_buffers(