Skip to content

Improve WebGPU MatMulNBits to support zero pointer for 2bits#27285

Merged
guschmue merged 3 commits intomicrosoft:mainfrom
HectorSVC:hecli_webgpu_2bit_zp
Feb 9, 2026
Merged

Improve WebGPU MatMulNBits to support zero pointer for 2bits#27285
guschmue merged 3 commits intomicrosoft:mainfrom
HectorSVC:hecli_webgpu_2bit_zp

Conversation

@HectorSVC
Copy link
Contributor

@HectorSVC HectorSVC commented Feb 9, 2026

Description

The existing WebGPU MatMulNBits op does not support zero pointer for 2bits. So it blocks some models. This PR enables the zero pointer support for 2bits support. UT tests are included for coverage.

@guschmue
Copy link
Contributor

guschmue commented Feb 9, 2026

/azp run Linux QNN CI Pipeline,Win_TRT_Minimal_CUDA_Test_CI,Windows ARM64 QNN CI Pipeline,Windows GPU Doc Gen CI Pipeline,Windows x64 QNN CI Pipeline

@guschmue guschmue added the ep:WebGPU ort-web webgpu provider label Feb 9, 2026
@azure-pipelines
Copy link

Azure Pipelines successfully started running 4 pipeline(s).

@guschmue guschmue merged commit 40d36b1 into microsoft:main Feb 9, 2026
88 checks passed
@HectorSVC HectorSVC deleted the hecli_webgpu_2bit_zp branch February 10, 2026 00:51
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

ep:WebGPU ort-web webgpu provider

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants