Use Phi-4-mini's own BOS and EOS ids in its example config

[john-rocky]

examples/models/phi_4_mini/config/phi_4_mini_xnnpack.yaml carries

metadata: '{"get_bos_id":151643, "get_eos_ids":[151643]}'

151643 is Qwen's <|endoftext|>. Phi-4-mini has a different vocabulary, and in it that id is a subword:

>>> from transformers import AutoTokenizer
>>> AutoTokenizer.from_pretrained("microsoft/Phi-4-mini-instruct").convert_ids_to_tokens(151643)
"Ġprzyp"
>>> AutoTokenizer.from_pretrained("Qwen/Qwen2.5-0.5B").convert_ids_to_tokens(151643)
"<|endoftext|>"

So the runner is told to begin on a fragment of a Polish word and to stop on one, and it never sees the token the model actually ends turns with.

microsoft/Phi-4-mini-instruct/generation_config.json gives:

{"bos_token_id": 199999, "eos_token_id": [200020, 199999]}

which are <|endoftext|> and <|end|>. This patch uses those.

Exported with the corrected ids (executorch 1.4.0, XNNPACK, 8da4w + 8-bit embedding, 2597.9 MB), the model answers and stops where it should:

prompt: <|user|>What is 17 times 4?<|end|><|assistant|>
output: 17 times 4 is equal to 68.
        decode: 10 tok in 0.20s (49.73 tok/s)

The generation ends after 10 tokens because <|end|> is reached, which is the part that does not happen with 151643.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/22042

• 📄 Preview Python docs built from this PR

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

⚠️ 20 Awaiting Approval

As of commit 3e14c95 with merge base e4576d0 ():

AWAITING APPROVAL - The following workflows need approval before CI can run:

• Build Aarch64 Linux CUDA Wheels (gh)
• Build Aarch64 Linux Wheels (gh)
• Build documentation (gh)
• Build Linux CUDA Wheels (gh)
• Build Linux Wheels (gh)
• Build macOS Wheels (gh)
• Build Presets (gh)
• Build Windows Wheels (gh)
• Lint (gh)
• periodic (gh)
• pull (gh)
• Test ARM Backend (gh)
• Test CoreML Backend (gh)
• Test Cortex-M Backend (gh)
• Test NXP Backend (gh)
• Test QNN Backend (gh)
• Test QNN Windows MSVC build (gh)
• Test Vulkan Backend (gh)
• Test WebGPU Backend (gh)
• Test XNNPACK Backend (gh)

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

[linux-foundation-easycla]

• ❌ The email address for the commit (3e14c95) is not linked to the GitHub account, preventing the EasyCLA check. Consult this Help Article and GitHub Help to resolve. (To view the commit's email address, add .patch at the end of this PR page's URL.) For further assistance with EasyCLA, please visit our EasyCLA portal and chat with our support bot.

[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.