Skip to content
build / sdist

[bugfix][torchair] fix wasted NPU memory buffer allocation for quantized deepseek with unquantized MTP layer #8289

Sign in to view logs

[bugfix][torchair] fix wasted NPU memory buffer allocation for quantized deepseek with unquantized MTP layer

[bugfix][torchair] fix wasted NPU memory buffer allocation for quantized deepseek with unquantized MTP layer #8289

Triggered via pull request September 20, 2025 14:12
@linfeng-yuanlinfeng-yuan
synchronize #3068
linfeng-yuan:fix_torchair_redundant_process_group
Status Success
Total duration 23s
Artifacts 1

release_code.yml

on: pull_request
Matrix: release code
Fit to window
Zoom out
Zoom in

Artifacts

Produced during runtime
Name Size Digest
vllm-ascend-src
1.28 MB
sha256:d9ce2a069d9bee77365e3c5bfb2d01f0fb8b39e66f187d549a7856dcc34c4ee1