Fix flaky testPhi4 and testVoxtral by setting temperature=0

[psiddh]

Summary:
Both tests were flaky because LLM outputs are non-deterministic with the default temperature of 0.8, which uses RNG-based sampling with a time-based seed. Setting temperature=0 enables greedy argmax decoding, eliminating randomness and making assertions on generated text reliable.

This is consistent with how other LLM tests and production runners in the codebase handle determinism (e.g., test_text_decoder_runner.cpp, test_sampler.cpp, and QNN/QAI Hub runners).

This fixes 5 flaky tests

{F1984490494}

Reviewed By: shoumikhin

Differential Revision: D90361187

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/16517

• 📄 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! (1 Unrelated Failure)

As of commit 8c9e574 with merge base 847d70d ():

UNSTABLE - The following job is marked as unstable, possibly due to flakiness on trunk:

• pull / android / run-emulator (gh) (#16137)
  Timeout waiting for emulator to boot.

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

[meta-codesync]

@psiddh has exported this pull request. If you are a Meta employee, you can view the originating Diff in D90361187.

[github-actions]

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

[Copilot]

Pull request overview

This PR fixes flaky LLM tests by setting temperature=0 in the generation configuration, enabling deterministic greedy decoding instead of random sampling. This eliminates non-deterministic behavior that was causing test failures when assertions checked for specific words in generated text.

Key changes:

• Added temperature=0 to both generation calls in testPhi4
• Added temperature=0 to the generation call in testVoxtral

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
extension/llm/apple/ExecuTorchLLM/tests/TextRunnerTest.swift	Set temperature=0 in both generation calls of testPhi4 to ensure deterministic output when asserting "paris" is in the generated text
extension/llm/apple/ExecuTorchLLM/tests/MultimodalRunnerTest.swift	Set temperature=0 in testVoxtral's generation call to ensure deterministic output when asserting "tattoo" is in the generated text

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.