Vald is a highly scalable distributed fast approximate nearest neighbor (ANN) dense vector search engine.
Vald is designed and implemented based on Cloud-Native architecture.
Vald has automatic vector indexing and index backup, and horizontal scaling which made for searching from billions of feature vector data.
Vald is easy to use, feature-rich and highly customizable as you needed.
It uses the fastest ANN Algorithm NGT to search neighbors.
(If you are interested in ANN benchmarks, please refer to ann-benchmarks.com.)
For more information, please refer to Official Web Site.
Vald can handle any object data, image, audio processing, video, text, binary, or etc., if converting to the vector, and be used for:
- Recognition
- Recommendation
- Detecting
- Grammar checker
- Real-time translator
- anything you want to do!
- Kubernetes 1.19~
- AVX2 instructions (required by Vald Agent NGT)
Go to Get Started page to try out Vald !
helm repo add vald https://vald.vdaas.org/charts
helm install vald-cluster vald/vald
If you use the default values.yaml, the nightly
images will be installed.
Please refer to vald-helm-operator.
Component | Docker image | latest image | nightly image |
---|---|---|---|
Agent NGT |
|
||
Agent Sidecar |
|
||
Discoverer |
|
||
Gateways |
|
|
|
Index Manager |
|
||
Helm Operator |
|
Docker images tagging policy:
nightly
... latest build of main branchvX.X.X
... released versionslatest
... latest build of release versionsstable
... latest long-term supported version
Please read the contribution guide.
Before your first commit to this repository, it is strongly recommended to run the commands below.
git clone https://github.com/vdaas/vald && cd vald
make init
Thanks goes to these wonderful people (emoji key):
Vald released under Apache 2.0 license, refer LICENSE file.