Hi there, second post of my personal writings to consolidade my understanding of things. As the first post, I tried to write it intuitively.
https://stillthinking.net/posts/connections-in-math-two-kinds-of-random/
Hi all, I have been now just writing things to consolidate some basic applied math. This is nothing advanced but just a good way to put things out and to learn by writing.
One of the things I try to do is to build things more intuitively instead of the traditional math book approach of starting from the final formalized result and giving a few formalized hints on how that object came from.
pynear is a fast KNN library that doesn't make you choose between exact answers and production speed. Drop-in for scikit-learn. SIMD-accelerated. Up to 257× faster than Faiss on binary descriptors at 100% recall.
• cosine indices (exact + approximate)
• drop-in for scikit-learn (metric='cosine')
• Incredibly fast approximate IndexBinaryMultiHash with high recall (beats Faiss in some contexts)
pip install -U pynear