Bump tokenizers to incorporate new features for HF tokenizer

[larryliu0820]

To incorporate changes in meta-pytorch/tokenizers#87, so that we can run gemma-3.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/11995

• 📄 Preview Python docs built from this PR

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

❗ 1 Active SEVs

There are 1 currently active SEVs. If your PR is affected, please view them below:

• VolumeLimitExceeded Issue for linux.2xlarge and linux.4xlarge

❌ 1 New Failure, 4 Unrelated Failures

As of commit 4750cc9 with merge base 478032a ():

NEW FAILURE - The following job has failed:

• pull / android / run-emulator (gh)
  The process '/usr/bin/sh' failed with exit code 255

BROKEN TRUNK - The following jobs failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

• pull / unittest / linux / linux-job (gh) (trunk failure)
  extension/llm/export/test/test_export_llm.py::TestExportLlm::test_with_config
• pull / unittest / macos / macos-job (gh) (trunk failure)
  extension/llm/export/test/test_export_llm.py::TestExportLlm::test_with_config
• pull / unittest-editable / linux / linux-job (gh) (trunk failure)
  extension/llm/export/test/test_export_llm.py::TestExportLlm::test_with_config
• pull / unittest-editable / macos / macos-job (gh) (trunk failure)
  extension/llm/export/test/test_export_llm.py::TestExportLlm::test_with_config

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

[guangy10]

🚀

[guangy10]

Benchmark scheduled to test against gemma-3

• iOS: https://github.com/pytorch/executorch/actions/runs/15908723568
• Android: https://github.com/pytorch/executorch/actions/runs/15908729469

[larryliu0820]

Perf benchmark jobs are passing. Merging this.