liuyu/pgvector
1
0
Fork 0
mirror of https://github.com/pgvector/pgvector.git synced 2026-07-11 07:06:57 +08:00
Code Issues Packages Projects Releases Wiki Activity

Updated readme [skip ci]

Browse Source
This commit is contained in:
Andrew Kane
2023-03-31 16:28:27 -07:00
parent 1d020abdd1
commit 5d0f88529e
1 changed files with 16 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
README.md
View File

@@ -173,24 +173,34 @@ CREATE TABLE items (embedding vector(3), category_id int) PARTITION BY LIST(cate
## Performance ## Performance
Use `EXPLAIN ANALYZE` to debug performance.
```sql
EXPLAIN ANALYZE SELECT * FROM items ORDER BY embedding <-> '[3,1,2]' LIMIT 1;
```
### Exact Search
To speed up queries without an index, increase `max_parallel_workers_per_gather`. To speed up queries without an index, increase `max_parallel_workers_per_gather`.
```sql ```sql
SET max_parallel_workers_per_gather = 4; SET max_parallel_workers_per_gather = 4;
``` ```
If vectors are normalized to length 1 (like those from [OpenAI](https://platform.openai.com/docs/guides/embeddings/limitations-risks)), use inner product instead of cosine distance for best performance.
```sql
SELECT * FROM items ORDER BY embedding <#> '[3,1,2]' LIMIT 1;
```
### Approximate Search
To speed up queries with an index, increase the number of inverted lists (at the expense of recall). To speed up queries with an index, increase the number of inverted lists (at the expense of recall).
```sql ```sql
CREATE INDEX ON items USING ivfflat (embedding vector_l2_ops) WITH (lists = 1000); CREATE INDEX ON items USING ivfflat (embedding vector_l2_ops) WITH (lists = 1000);
``` ```
Use `EXPLAIN ANALYZE` to debug performance.
```sql
EXPLAIN ANALYZE SELECT * FROM items ORDER BY embedding <-> '[3,1,2]' LIMIT 1;
```
## Languages ## Languages
Use pgvector from any language with a Postgres client. You can even generate and store vectors in one language and query them in another. Use pgvector from any language with a Postgres client. You can even generate and store vectors in one language and query them in another.