Chranos/enginex-mlu370-vllm
1
0
Fork 0
forked from EngineX-Cambricon/enginex-mlu370-vllm
Code Pull Requests Activity

add deepseekv3 and llama4

Browse Source
This commit is contained in:
Chranos
2026-02-11 15:48:35 +08:00
parent e752946445
commit 6eae065dd6
1 changed files with 3 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
vllm-v0.6.2/vllm/model_executor/layers/vocab_parallel_embedding.py
View File

@@ -38,6 +38,9 @@ class UnquantizedEmbeddingMethod(QuantizeMethodBase):
layer: torch.nn.Module, layer: torch.nn.Module,
x: torch.Tensor, x: torch.Tensor,
bias: Optional[torch.Tensor] = None) -> torch.Tensor: bias: Optional[torch.Tensor] = None) -> torch.Tensor:
# MLU F.linear requires matching dtypes
if x.dtype != layer.weight.dtype:
x = x.to(layer.weight.dtype)
return F.linear(x, layer.weight, bias) return F.linear(x, layer.weight, bias)
def embedding(self, layer: torch.nn.Module, def embedding(self, layer: torch.nn.Module,