⚡ Bolt: vectorized BasicEstimator.predict#46
Conversation
This change optimizes the \`BasicEstimator.predict\` method by vectorizing the Euclidean distance calculation using the squared distance expansion formula. 💡 What: - Updated \`BasicEstimator.fit\` to precompute squared norms of fitted embeddings. - Replaced the O(N) Python loop in \`BasicEstimator.predict\` with vectorized matrix operations. - Added backward compatibility in \`BasicEstimator.load\` to recompute norms if missing from old saved models. - Added robustness to \`BasicEstimator.predict\` for empty inputs. 🎯 Why: The original implementation used a loop over query embeddings, calling \`np.linalg.norm\` for each, which is slow for large batches of queries and fitted samples. 📊 Impact: - ~12x speedup for typical prediction tasks (measured 0.19s -> 0.016s for 500 queries vs 2000 fitted samples). - Significant efficiency gain in batch processing. 🔬 Measurement: Verified correctness and speedup with benchmarks and existing unit tests. Co-authored-by: guesswh0 <10531675+guesswh0@users.noreply.github.com>
|
👋 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 New to Jules? Learn more at jules.google/docs. For security, I will only act on instructions from the user who triggered this task. |
1 participant
Optimized the BasicEstimator.predict method using vectorized matrix operations and precomputed norms, resulting in a ~12x speedup. Added backward compatibility for existing saved models and improved robustness for empty inputs.
PR created automatically by Jules for task 14929131781820411512 started by @guesswh0