[LLAVA] Enable 2nd XNNPACK Partition pass for the text model

[digantdesai]

This is to pick up ops like mul, add, sigmoid etc. which contributes to ghe e2e latency.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/4968

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

✅ You can merge normally! (2 Unrelated Failures)

As of commit dc164f5 with merge base 7b3549b ():

FLAKY - The following jobs failed but were likely due to flakiness present on trunk:

• pull / test-selective-build-linux (buck2) / linux-job (gh) (detected as infra flaky with no log or failing log classifier)
• pull / unittest-arm (buck2) / linux-job (gh) (detected as infra flaky with no log or failing log classifier)

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[larryliu0820]

Just tested on my OnePlus 9 pro (12 GB RAM). Time to first token reduced from 35 seconds to 30 seconds.

[facebook-github-bot]

@digantdesai has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.