-
Hi, I wonder which model is used on the embedding vectors provided for the 100M dataset?
Thanks in advance!
-
### Description
Currently, if you want to use PostgreSQL with the `hibernate-vector` dependency, when using DevServices with PostgreSQL database, it starts the service with a standard PostgreSQL data…
-
When trying to get the same result from a keras 2 and keras 3 models I noticed that there were some differences so I investigated further with this test that failed on the assertion:
```
def t…
-
### Environment
docker
image tensorchord/pgvecto-rs:pg16-v0.3.0
image tensorchord/pgvecto-rs: pg16-v0.2.1
### Step
1. create table with embedding vector(3072)
2. create index…
-
### Description
Some inference API providers now support embedding models with each dimension defined as a single bit. For example, the v3 models from Cohere offer this capability. Since we already h…
-
**The problem/use-case that the feature addresses**
Enable use of Valkey with LLM applications for semantic LLM caching, semantic conversation cache, LLM semantic routing
**Description of the fe…
-
The PgVectorEmbeddingStore enables query operations that exclusively leverage metadata for filtering criteria, without mandating the inclusion of embedding vectors, minimum similarity scores (minScore…
-
Multi-vector MaxSim is increasingly important and we have optimizations for float cell precision, but I think we should also consider optimize for int8 with hamming as it approximates the dotproduct f…
-
-
Hi @tamerthamoqa ,
Thanks a lot for your great repo.
According to FaceNet paper the best dimension for embedded vector is 128. I am curious to know is there any specific reason that you used four …