[cherry-pick]Upgrade CANN to 8.3.rc1 (#3945) (#3962)

This PR upgrade CANN from 8.2rc1 to 8.3rc1 and remove the CANN version
check logic.

TODO: we notice that UT runs failed with CANN 8.3 image. So the base
image for UT is still 8.2. We'll fix it later.

- vLLM version: v0.11.0
- vLLM main:
83f478bb19

Signed-off-by: wangxiyuan <wangxiyuan1007@gmail.com>
This commit is contained in:
wangxiyuan
2025-11-06 09:05:08 +08:00
committed by GitHub
parent 66b67f9cf2
commit 7ee0b0b5d8
36 changed files with 104 additions and 192 deletions

View File

@@ -45,8 +45,8 @@ class AscendUnquantizedLinearMethod(UnquantizedLinearMethod):
def process_weights_after_loading(self, layer: torch.nn.Module) -> None:
super().process_weights_after_loading(layer)
if (is_enable_nz() and torch.version.cann.startswith("8.3") and
layer.weight.data.dtype in [torch.float16, torch.bfloat16]):
if (is_enable_nz() and layer.weight.data.dtype
in [torch.float16, torch.bfloat16]):
layer.weight.data = torch_npu.npu_format_cast(
layer.weight.data, ACL_FORMAT_FRACTAL_NZ)