Releases: milvus-io/knowhere
knowhere-v1.3.17
Release : knowhere-v1.3.17
What's Changed
- Add new interface HasRawData (#947)
- HasRawData() returns True for HNSW only, and let GetVectorByIds() gets dim from index data (#957)
Full Changelog: v1.3.16...v1.3.17
knowhere-v1.3.16
Release : knowhere-v1.3.16
What's Changed
- For
HNSW
, the range ofef
is changed tok
tomax
. Andef
is not required to be in the config anymore, default value will bemax(k, 16)
. - For
DiskANN
, the range ofsearch_list
is now changed tok
tomax
. Alsosearch_list
doesn't have to be in the config, and a default value ofmax(k, 16)
will be set.
Full Changelog: v1.3.15...v1.3.16
knowhere-v1.3.15
Release Note
Enhancement
- Set Build & Search default params for all index types except Annoy (#908 #909)
- Modify the search list size if this value not set in struct or too small (#911)
Full Changelog: v1.3.14...v1.3.15
knowhere-v1.3.14
Release Note
Knowhere-1.3.14
Knowhere-v1.3.14 is a bug fix version solving some existed issues.
Bug Fix
- Set DiskANN filter threshold to default value when missing (#897)
- Fix the lower recall issue in DiskANN when using metric IP with unnormalized data (#904)
Full Changelog: v1.3.13...v1.3.14
knowhere-v1.3.13
Release Note
Knowhere-1.3.13
- Added LRU cache for the entry point of HNSW and DiskANN to speed up similar requests. #843
- Used Brute Force in HNSW to improve performance when the filter rate is too high. #861
- Used PQ Based Brute Force + Refine in DiskANN when the filter rate is too high. #861
Bug fix
- This fix resolves the issue where DiskANN is unable to locate the first vector of an index #854.
Full Changelog: v1.3.12...v1.3.13
knowhere-v1.3.12
Release : knowhere-v1.3.12
What's Changed
- Fix DiskANN Access Invalid Address when Query with Bitset (#798) by @PwzXxm in #806
- Change the Ef limit for HNSW when topk is large by @xiaofan-luan in #809
Full Changelog: v1.3.11...v1.3.12
knowhere-v1.3.11
Release Note
Knowhere-1.3.11
Knowhere v1.3.11 is a bug fix version solving some exited issues.
Enhancement
Bug Fix
Full Changelog: v1.3.10...v1.3.11
Knowhere 2.1.0 Beta
Release Note
Knowhere-2.1.0 Beta
We are happy to announce the release of Knowhere 2.1.0 Beta! Knowhere 2.1.0 newly supports RAFT GPU-based IVF index and FAISS GPU-based IVF index.
Special Thanks
@wphicks @cjnolet from NVIDIA contributed the RAFT integration code!
Breaking Change
- Support FAISS GPU-based IVF-series indexes. (#708)
- Support RAFT GPU-based IVF-series indexes. (#712, #734)
RAFT is an open-source library for GPU-based ANN search, developed by Nvidia. It offers impressive performance and a dependable GPU-based ANN index. With Knowhere 2.1.0, Nvidia has incorporated its advanced GPU solutions to enhance Knowhere's capabilities.
We conducted benchmark tests on RAFT's GPU-based IVF-series indexes and achieved remarkable results. All benchmarks are conducted in a host of 8c32g, and an Nvidia A100 GPU. NQ in these benchmarks are 100.
Benchmark1
First we did some comparison between RAFT-IVF-Flat
with IVF-Flat
and HNSW
at 95% recall.
DataSets | SIFT | GIST | GLOVE | DEEP |
---|---|---|---|---|
HNSW (VPS) | 14,537 | 791 | 1,516 | 5,761 |
IVF-Flat (VPS) | 3,097 | 142 | 791 | 723 |
RAFT-IVF-Flat (VPS) | 121,568 | 5,737 | 20,163 | 16,557 |
Benchmark2
Also we benchmarked RAFT-IVF-PQ
comparing Knowhere's fastest index HNSW
at 80% recall.
DataSets | SIFT | GIST | GLOVE | DEEP |
---|---|---|---|---|
HNSW (VPS) | 20,809 | 2,593 | 8,005 | 13,291 |
RAFT-IVF-PQ (VPS) | 271,885 | 7,448 | 38,989 | 80,363 |
Enhancement
- Improve performance of CPU-based IVF-series and BruteForce. (#703)
- Related Issue: milvus-io/milvus#22163
Bug fix
- Fix TANIMOTO distance calculation error in BruteFoce. (#719)
New Contributors
knowhere-v1.3.1-hotfix
Release : knowhere-v1.3.1-hotfix
knowhere-v1.3.10
Release : knowhere-v1.3.10