Skip to content

⚡ Bolt: Vectorize BasicEstimator.predict#21

Open
guesswh0 wants to merge 1 commit into
masterfrom
bolt-vectorize-basic-estimator-4169103200251108033
Open

⚡ Bolt: Vectorize BasicEstimator.predict#21
guesswh0 wants to merge 1 commit into
masterfrom
bolt-vectorize-basic-estimator-4169103200251108033

Conversation

@guesswh0
Copy link
Copy Markdown
Owner

@guesswh0 guesswh0 commented May 1, 2026

Vectorized BasicEstimator.predict for significant performance gains in face recognition tasks.


PR created automatically by Jules for task 4169103200251108033 started by @guesswh0

💡 What:
Replaced the iterative Euclidean distance calculation in `BasicEstimator.predict` with a fully vectorized implementation using the squared distance expansion formula $\|a-b\|^2 = \|a\|^2 + \|b\|^2 - 2a \cdot b$.
Also updated `BasicEstimator.fit` to pre-calculate and store the squared norms of the fitted embeddings (`self.norms_sq`) for additional speedup.

🎯 Why:
The original implementation used a Python-level loop over input embeddings and called `np.linalg.norm` for each, which is inefficient for large numbers of samples. Vectorization allows NumPy to use optimized BLAS routines, significantly reducing computation time.

📊 Impact:
Measurable performance improvement for batch predictions. In local benchmarks (500 query samples against 2000 fitted samples), prediction time dropped from ~0.26s to ~0.05s (approx. 5x speedup).

🔬 Measurement:
Run a benchmark comparing the original and new `predict` methods with a large number of embeddings. Functional correctness is verified by the existing test suite (`python3 -m unittest discover tests`).

Backward compatibility is maintained for models fitted with older versions by using `getattr(self, "norms_sq", None)` as a fallback. Floating-point noise is handled via `np.maximum(..., 0)`.

Co-authored-by: guesswh0 <10531675+guesswh0@users.noreply.github.com>
@google-labs-jules
Copy link
Copy Markdown
Contributor

👋 Jules, reporting for duty! I'm here to lend a hand with this pull request.

When you start a review, I'll add a 👀 emoji to each comment to let you know I've read it. I'll focus on feedback directed at me and will do my best to stay out of conversations between you and other bots or reviewers to keep the noise down.

I'll push a commit with your requested changes shortly after. Please note there might be a delay between these steps, but rest assured I'm on the job!

For more direct control, you can switch me to Reactive Mode. When this mode is on, I will only act on comments where you specifically mention me with @jules. You can find this option in the Pull Request section of your global Jules UI settings. You can always switch back!

New to Jules? Learn more at jules.google/docs.


For security, I will only act on instructions from the user who triggered this task.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant