Sfrgttk

Multiplying elements of a list

Coworker asking me to not bring cakes due to self control issue. What should I do?

How do I write a maintainable, fast, compile-time bit-mask in C++?

Can a Hydra make multiple opportunity attacks at once?

How to achieve physical gender equality?

SQL Server 2017 crashes when backing up because filepath is wrong

Manager has noticed coworker's excessive breaks. Should I warn him?

Dot product with a constant

How can changes in personality/values of a person who turned into a vampire be explained?

Was the Soviet N1 really capable of sending 9.6 GB/s of telemetry?

Is Screenshot Time-tracking Common?

How can guns be countered by melee combat without raw-ability or exceptional explanations?

Multiple null checks in Java 8

How do I add a strong "onion flavor" to the biryani (in restaurant style)?

Is it possible to detect 100% of SQLi with a simple regex?

Is there a way to pause a running process on Linux systems and resume later?

Have the UK Conservatives lost the working majority and if so, what does this mean?

Does this property characterize the odd primes?

Given the mapping a = 1, b = 2, ... z = 26, and an encoded message, count the number of ways it can be decoded

Is it common to refer to someone as "Prof. Dr. [LastName]"?

Have any astronauts or cosmonauts died in space?

For the Circle of Spores druid's Halo of Spores feature, is your reaction used regardless of whether the other creature succeeds on the saving throw?

Are Kickstages exclusive to Rocketlabs Electron Design?

Why would you use 2 alternate layout buttons instead of 1, when only one can be selected at once

Postgresql: How to store word embedding vectors and lookup by cosine similarity?

Quick nearest neighbor search in the 150-dimensional spaceUsing an index for both uniqueness and fast lookup?PostgreSQL FTS and Trigram-similarity Query OptimizationPostgreSQL return word after matching first characterDesign options for time series scientific dataImproving query performance involving similarity and full-text searchHow does postgres store array values?PostgreSQL: how can I shorten a multi-word string to few words?How exactly does trigram word-similarity work?PostgreSQL Update Column From LookupHow can I index a text[] field in postgresql?

0

I am trying to store vectors for word/doc embeddings in a postgresql table, and want to be able to quickly pull the N rows with highest cosine similarity to a given query vector. The vectors I'm working with are numpy.arrays of floats with length 100 <= L <= 1000.

I looked into the cube module for similarity search, but it is limited to vectors with <= 100 dimensions. The embeddings I am using will result in vectors that are 100-dimensions minimum and often much higher (depending on settings when training word2vec/doc2vec models).

What is the most efficient way to store large dimensional vectors (numpy float arrays) in postgres, and perform quick lookup based on cosine similarity (or other vector similarity metrics)?

postgresql index array dimension

share

asked 4 mins ago

J. TaylorJ. Taylor

132213

add a comment | 

0

I am trying to store vectors for word/doc embeddings in a postgresql table, and want to be able to quickly pull the N rows with highest cosine similarity to a given query vector. The vectors I'm working with are numpy.arrays of floats with length 100 <= L <= 1000.

I looked into the cube module for similarity search, but it is limited to vectors with <= 100 dimensions. The embeddings I am using will result in vectors that are 100-dimensions minimum and often much higher (depending on settings when training word2vec/doc2vec models).

What is the most efficient way to store large dimensional vectors (numpy float arrays) in postgres, and perform quick lookup based on cosine similarity (or other vector similarity metrics)?

postgresql index array dimension

share

asked 4 mins ago

J. TaylorJ. Taylor

132213

add a comment | 

I am trying to store vectors for word/doc embeddings in a postgresql table, and want to be able to quickly pull the N rows with highest cosine similarity to a given query vector. The vectors I'm working with are numpy.arrays of floats with length 100 <= L <= 1000.

I looked into the cube module for similarity search, but it is limited to vectors with <= 100 dimensions. The embeddings I am using will result in vectors that are 100-dimensions minimum and often much higher (depending on settings when training word2vec/doc2vec models).

What is the most efficient way to store large dimensional vectors (numpy float arrays) in postgres, and perform quick lookup based on cosine similarity (or other vector similarity metrics)?

postgresql index array dimension

share

asked 4 mins ago

J. TaylorJ. Taylor

132213

share

asked 4 mins ago

J. TaylorJ. Taylor

132213

share

asked 4 mins ago

J. TaylorJ. Taylor

132213

asked 4 mins ago

J. TaylorJ. Taylor

132213

asked 4 mins ago

J. TaylorJ. Taylor

132213