liuyu/pgvector
1
0
Fork 0
mirror of https://github.com/pgvector/pgvector.git synced 2026-07-15 00:56:54 +08:00
Code Issues Packages Projects Releases Wiki Activity

Updated readme [skip ci]

Browse Source
This commit is contained in:
Andrew Kane
2025-02-20 00:05:35 -08:00
parent 78ed8f1157
commit fef635c9e5
1 changed files with 21 additions and 61 deletions
Download Patch File Download Diff File Expand all files Collapse all files

82
README.md
View File

@@ -82,7 +82,7 @@ Get the nearest neighbors by L2 distance
SELECT * FROM items ORDER BY embedding <-> '[3,1,2]' LIMIT 5; SELECT * FROM items ORDER BY embedding <-> '[3,1,2]' LIMIT 5;
``` ```
Also supports inner product (`<#>`), cosine distance (`<=>`), and L1 distance (`<+>`, added in 0.7.0) Also supports inner product (`<#>`), cosine distance (`<=>`), and L1 distance (`<+>`)
Note: `<#>` returns the negative inner product since Postgres only supports `ASC` order index scans on operators Note: `<#>` returns the negative inner product since Postgres only supports `ASC` order index scans on operators
@@ -146,9 +146,9 @@ Supported distance functions are:
- `<->` - L2 distance - `<->` - L2 distance
- `<#>` - (negative) inner product - `<#>` - (negative) inner product
- `<=>` - cosine distance - `<=>` - cosine distance
- `<+>` - L1 distance (added in 0.7.0) - `<+>` - L1 distance
- `<~>` - Hamming distance (binary vectors, added in 0.7.0) - `<~>` - Hamming distance (binary vectors)
- `<%>` - Jaccard distance (binary vectors, added in 0.7.0) - `<%>` - Jaccard distance (binary vectors)
Get the nearest neighbors to a row Get the nearest neighbors to a row
@@ -235,19 +235,19 @@ Cosine distance
CREATE INDEX ON items USING hnsw (embedding vector_cosine_ops); CREATE INDEX ON items USING hnsw (embedding vector_cosine_ops);
``` ```
L1 distance - added in 0.7.0 L1 distance
```sql ```sql
CREATE INDEX ON items USING hnsw (embedding vector_l1_ops); CREATE INDEX ON items USING hnsw (embedding vector_l1_ops);
``` ```
Hamming distance - added in 0.7.0 Hamming distance
```sql ```sql
CREATE INDEX ON items USING hnsw (embedding bit_hamming_ops); CREATE INDEX ON items USING hnsw (embedding bit_hamming_ops);
``` ```
Jaccard distance - added in 0.7.0 Jaccard distance
```sql ```sql
CREATE INDEX ON items USING hnsw (embedding bit_jaccard_ops); CREATE INDEX ON items USING hnsw (embedding bit_jaccard_ops);
@@ -256,9 +256,9 @@ CREATE INDEX ON items USING hnsw (embedding bit_jaccard_ops);
Supported types are: Supported types are:
- `vector` - up to 2,000 dimensions - `vector` - up to 2,000 dimensions
- `halfvec` - up to 4,000 dimensions (added in 0.7.0) - `halfvec` - up to 4,000 dimensions
- `bit` - up to 64,000 dimensions (added in 0.7.0) - `bit` - up to 64,000 dimensions
- `sparsevec` - up to 1,000 non-zero elements (added in 0.7.0) - `sparsevec` - up to 1,000 non-zero elements
### Index Options ### Index Options
@@ -312,7 +312,7 @@ Note: Do not set `maintenance_work_mem` so high that it exhausts the memory on t
Like other index types, it’s faster to create an index after loading your initial data Like other index types, it’s faster to create an index after loading your initial data
Starting with 0.6.0, you can also speed up index creation by increasing the number of parallel workers (2 by default) You can also speed up index creation by increasing the number of parallel workers (2 by default)
```sql ```sql
SET max_parallel_maintenance_workers = 7; -- plus leader SET max_parallel_maintenance_workers = 7; -- plus leader
@@ -365,7 +365,7 @@ Cosine distance
CREATE INDEX ON items USING ivfflat (embedding vector_cosine_ops) WITH (lists = 100); CREATE INDEX ON items USING ivfflat (embedding vector_cosine_ops) WITH (lists = 100);
``` ```
Hamming distance - added in 0.7.0 Hamming distance
```sql ```sql
CREATE INDEX ON items USING ivfflat (embedding bit_hamming_ops) WITH (lists = 100); CREATE INDEX ON items USING ivfflat (embedding bit_hamming_ops) WITH (lists = 100);
@@ -374,8 +374,8 @@ CREATE INDEX ON items USING ivfflat (embedding bit_hamming_ops) WITH (lists = 10
Supported types are: Supported types are:
- `vector` - up to 2,000 dimensions - `vector` - up to 2,000 dimensions
- `halfvec` - up to 4,000 dimensions (added in 0.7.0) - `halfvec` - up to 4,000 dimensions
- `bit` - up to 64,000 dimensions (added in 0.7.0) - `bit` - up to 64,000 dimensions
### Query Options ### Query Options
@@ -547,8 +547,6 @@ Note: If this is lower than `ivfflat.probes`, `ivfflat.probes` will be used
## Half-Precision Vectors ## Half-Precision Vectors
*Added in 0.7.0*
Use the `halfvec` type to store half-precision vectors Use the `halfvec` type to store half-precision vectors
```sql ```sql
@@ -557,8 +555,6 @@ CREATE TABLE items (id bigserial PRIMARY KEY, embedding halfvec(3));
## Half-Precision Indexing ## Half-Precision Indexing
*Added in 0.7.0*
Index vectors at half precision for smaller indexes Index vectors at half precision for smaller indexes
```sql ```sql
@@ -580,24 +576,16 @@ CREATE TABLE items (id bigserial PRIMARY KEY, embedding bit(3));
INSERT INTO items (embedding) VALUES ('000'), ('111'); INSERT INTO items (embedding) VALUES ('000'), ('111');
``` ```
Get the nearest neighbors by Hamming distance (added in 0.7.0) Get the nearest neighbors by Hamming distance
```sql ```sql
SELECT * FROM items ORDER BY embedding <~> '101' LIMIT 5; SELECT * FROM items ORDER BY embedding <~> '101' LIMIT 5;
``` ```
Or (before 0.7.0)
```sql
SELECT * FROM items ORDER BY bit_count(embedding # '101') LIMIT 5;
```
Also supports Jaccard distance (`<%>`) Also supports Jaccard distance (`<%>`)
## Binary Quantization ## Binary Quantization
*Added in 0.7.0*
Use expression indexing for binary quantization Use expression indexing for binary quantization
```sql ```sql
@@ -620,8 +608,6 @@ SELECT * FROM (
## Sparse Vectors ## Sparse Vectors
*Added in 0.7.0*
Use the `sparsevec` type to store sparse vectors Use the `sparsevec` type to store sparse vectors
```sql ```sql
@@ -655,8 +641,6 @@ You can use [Reciprocal Rank Fusion](https://github.com/pgvector/pgvector-python
## Indexing Subvectors ## Indexing Subvectors
*Added in 0.7.0*
Use expression indexing to index subvectors Use expression indexing to index subvectors
```sql ```sql
@@ -1171,6 +1155,12 @@ cd pgvector
docker build --pull --build-arg PG_MAJOR=17 -t myuser/pgvector . docker build --pull --build-arg PG_MAJOR=17 -t myuser/pgvector .
``` ```
If you increase `maintenance_work_mem`, make sure `--shm-size` is at least that size to avoid an error with parallel HNSW index builds.
```sh
docker run --shm-size=1g ...
```
### Homebrew ### Homebrew
With Homebrew Postgres, you can use: With Homebrew Postgres, you can use:
@@ -1258,36 +1248,6 @@ You can check the version in the current database with:
SELECT extversion FROM pg_extension WHERE extname = 'vector'; SELECT extversion FROM pg_extension WHERE extname = 'vector';
``` ```
## Upgrade Notes
### 0.6.0
#### Postgres 12
If upgrading with Postgres 12, remove this line from `sql/vector--0.5.1--0.6.0.sql`:
```sql
ALTER TYPE vector SET (STORAGE = external);
```
Then run `make install` and `ALTER EXTENSION vector UPDATE;`.
#### Docker
The Docker image is now published in the `pgvector` org, and there are tags for each supported version of Postgres (rather than a `latest` tag).
```sh
docker pull pgvector/pgvector:pg16
# or
docker pull pgvector/pgvector:0.6.0-pg16
```
Also, if you’ve increased `maintenance_work_mem`, make sure `--shm-size` is at least that size to avoid an error with parallel HNSW index builds.
```sh
docker run --shm-size=1g ...
```
## Thanks ## Thanks
Thanks to: Thanks to: