More representative runner with better timing

[mergennachin]

DYLD_LIBRARY_PATH=/usr/lib ./cmake-out/examples/models/parakeet/parakeet_runner \
  --model_path examples/models/parakeet/parakeet_metal_optimized_argmax/parakeet_tdt.pte \
  --data_path examples/models/parakeet/parakeet_metal_optimized_argmax/aoti_metal_blob.ptd \
  --audio_path examples/models/parakeet/output.wav \
  --tokenizer_path examples/models/parakeet/parakeet_metal_optimized_argmax/tokenizer.model \
  --repeat 5 --warmup_repeat 3

Here's the output on M1 laptop

https://gist.github.com/mergennachin/a0ffb2b95bb398031ab420e4ec60f499

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/16641

• 📄 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 6534ac7 with merge base d58c8ee ():

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

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

• pull / android / run-emulator (gh) (trunk failure)
  The process '/usr/bin/sh' failed with exit code 255
• pull / test-samsung-models-linux / linux-job (gh) (trunk failure)
  ##[error]The operation was canceled.

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

[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 refactors the Parakeet model runner to improve timing instrumentation and make it more representative for benchmarking. The changes introduce a ParakeetRunner class that encapsulates model loading and inference, adds granular timing measurements for different pipeline stages, and supports warmup and repeated inference runs.

Changes:

• Introduced ParakeetRunner class to encapsulate model, tokenizer, and configuration
• Added detailed timing instrumentation (preprocessor, encoder, decoder components)
• Added --repeat and --warmup_repeat flags for benchmarking

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
examples/models/parakeet/main.cpp	Refactored to use ParakeetRunner class with timing instrumentation and benchmarking support
examples/models/parakeet/README.md	Updated documentation for new flags and corrected default tokenizer path

---

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

[Copilot]

Using std::max(1, ...) ensures values are at least 1, but FLAGS_repeat and FLAGS_warmup_repeat are already defined with default value 1. If the intent is to guard against negative values, consider adding validation with clear error messages instead of silently clamping to 1. This would make invalid input more obvious to users.

Suggested change

	const int num_iterations = std::max(1, FLAGS_repeat);
	const int num_warmup = std::max(1, FLAGS_warmup_repeat);
	if (FLAGS_repeat <= 0) {
	ET_LOG(
	Error,
	"Invalid value for --repeat: %d. Expected a positive integer.",
	FLAGS_repeat);
	return 1;
	}
	if (FLAGS_warmup_repeat <= 0) {
	ET_LOG(
	Error,
	"Invalid value for --warmup_repeat: %d. Expected a positive integer.",
	FLAGS_warmup_repeat);
	return 1;
	}
	const int num_iterations = FLAGS_repeat;
	const int num_warmup = FLAGS_warmup_repeat;

[JacobSzwejbka]

reinventing asr runner from first principals lol

[JacobSzwejbka]

Is this just defensive llm stuff