A post introducing Evoke claims its top-100 recall matches a 0.6B dense embedding model's to within 0.004.
A post introducing Evoke on October 8 says the 30M-parameter model is open source and runs on CPUs inside Postgres. It claims Evoke's top-100 recall matches a 0.6B dense embedding model's to within 0.004 and that Evoke finds more in the top 1,000.
36.6K
4 Sources, first seen ago
—
Not ranked yet
—
Not ranked yet
@ii_postsA 30M-parameter model, on CPUs, inside Postgres It matches a 0.6B dense embedding model on top-100 recall, to within 0.004, and finds more in the top 1,000 Meet Evoke. Open source today.
@EMostaqueRT @ii_posts: A 30M-parameter model, on CPUs, inside Postgres It matches a 0.6B dense embedding model on top-100 recall, to within 0.004,…@ii_postsA 30M-parameter model, on CPUs, inside Postgres It matches a 0.6B dense embedding model on top-100 recall, to within 0.004, and finds more in the top 1,000 Meet Evoke. Open source today.@EMostaqueRT @ii_posts: A 30M-parameter model, on CPUs, inside Postgres It matches a 0.6B dense embedding model on top-100 recall, to within 0.004,…@ii_postsA 30M-parameter model, on CPUs, inside Postgres It matches a 0.6B dense embedding model on top-100 recall, to within 0.004, and finds more in the top 1,000 Meet Evoke. Open source today.@EMostaqueRT @ii_posts: A 30M-parameter model, on CPUs, inside Postgres It matches a 0.6B dense embedding model on top-100 recall, to within 0.004,…@ii_postsA 30M-parameter model, on CPUs, inside Postgres It matches a 0.6B dense embedding model on top-100 recall, to within 0.004, and finds more in the top 1,000 Meet Evoke. Open source today.@EMostaqueRT @ii_posts: A 30M-parameter model, on CPUs, inside Postgres It matches a 0.6B dense embedding model on top-100 recall, to within 0.004,…
