3 ms·
Nice work! I've made a similar demo as a web app [1] and obtained a few orders of magnitude faster performance, in the 50ms for 1.6M images on single laptop cor
by m3at 4y ago
Nice work! I've made a similar demo as a web app [1] and obtained a few orders of magnitude faster performance, in the 50ms for 1.6M images on single laptop core. I've limited experience with coreml but there's a few tricks that could probably help you too? If you feel like hacking on it together shoot me an email :) (in profile)
[1] https://paulw.tokyo/post/real-time-semantic-search-demo/ https://paulw.tokyo/post/real-time-semantic-search-demo/
- mazzystar 4y agoThank you, I took a brief look and it's fantastic! I'll take a closer look tomorrow! (It's late at night here.) Would love to work with you if possible :)
- ntonozzi 4y agoWhat did you use as the index structure for the nearest neighbor search? Did you use a library like FAISS or Annoy?
- mazzystar 4y agoNo, just calculate the cosine similarity one by one in the simplest way.