Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[NPUW] extend DQ & PMM processing and make reduceSum not to keep axis #26779

Merged
merged 8 commits into from
Oct 15, 2024

Conversation

shaojun-yao
Copy link
Contributor

@shaojun-yao shaojun-yao commented Sep 25, 2024

Details:

  • extend DQ and PMM to support more patterns. e.g. fp16 matmuls
  • Make reduceSum not to keep axis because then it will convert to poolings in compiler. Otherwise reduceSum will convert to the convolution which is less efficient than poolings.

Tickets:

  • E-140570

@shaojun-yao shaojun-yao requested review from a team as code owners September 25, 2024 07:49
@github-actions github-actions bot added category: NPU OpenVINO NPU plugin category: NPUW NPUW plugin labels Sep 25, 2024
@sys-openvino-ci sys-openvino-ci added the ExternalIntelPR External contributor from Intel label Sep 25, 2024
@shaojun-yao shaojun-yao changed the title [draft] Sj/extend d qand pmm [NPUW] extend DQ & PMM processing and make reduceSum not to keep axis Sep 25, 2024
@dmatveev dmatveev self-assigned this Sep 25, 2024
@dmatveev dmatveev added this to the 2024.5 milestone Sep 25, 2024
@dmatveev dmatveev enabled auto-merge October 11, 2024 15:12
@sbutnari
Copy link
Contributor

build_jenkins

@dmatveev dmatveev added this pull request to the merge queue Oct 15, 2024
Merged via the queue into openvinotoolkit:master with commit c6801aa Oct 15, 2024
135 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
category: NPU OpenVINO NPU plugin category: NPUW NPUW plugin ExternalIntelPR External contributor from Intel
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants