Skip to content

fix: use v1.0.0 Gemma Omni Vulkan model - #1628

Open
barhanc wants to merge 1 commit into
mainfrom
@bh/investigate-1613-gemma-vulkan-memory
Open

barhanc wants to merge 1 commit into
mainfrom
@bh/investigate-1613-gemma-vulkan-memory

Conversation

@barhanc

@barhanc barhanc commented Nov 9, 2026 •

Copy link
Copy Markdown
Member

Description

Fixes the memory issue with Gemma Omni vulkan export. See: software-mansion-labs/executorch-export-recipes#11

Introduces a breaking change?

  • Yes
  • No

Type of change

  • Bug fix (change which fixes an issue)
  • New feature (change which adds functionality)
  • Documentation update (improves or adds clarity to existing documentation)
  • Other (chores, tests, code style improvements etc.)

Tested on

  • iOS
  • Android

Testing instructions

  • NLP example app -> test Vulkan Gemma Omni (ideally on 8GB Android phone e.g. S24)

Screenshots

Related issues

Fixes #1613

Checklist

  • I have performed a self-review of my code
  • I have commented my code, particularly in hard-to-understand areas
  • I have updated the documentation accordingly
  • My changes generate no new warnings

Additional notes

@barhanc barhanc self-assigned this Nov 9, 2026
@barhanc barhanc added model Issues related to exporting, improving, fixing ML models bug fix PRs that are fixing bugs labels Nov 9, 2026
@barhanc
barhanc requested a review from msluszniak November 16, 2026 17:02

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

bug fix PRs that are fixing bugs model Issues related to exporting, improving, fixing ML models

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Gemma 4 VL (Vulkan, v0.11.0 export) is killed by lmkd during load on an 8 GB phone; the v0.9.0 export loads there through the legacy runner

1 participant