Releases: milvus-io/milvus-model
Releases · milvus-io/milvus-model
v0.2.11
v0.2.10
What's Changed
- bgem3 devices kwargs syntax error bugfix by @SamSweere in #50
- fix: Change BGEM3 parameter to 'devices'. by @wxywb in #51
New Contributors
- @SamSweere made their first contribution in #50
Full Changelog: v0.2.9...v0.2.10
v0.2.9
What's Changed
- Fix: Correct parameter name in _encode method for BGEM3FlagModel by @thititj in #45
- fix: Adapt bge reranker and bge m3 to FlagEmbedding>1.3.0 by @wxywb in #47
- fix: Set default device in BGEM3EmbeddingFunction to None. by @wxywb in #48
Summary
- Adapt to FlagEmbedding > 1.3
New Contributors
Full Changelog: v0.2.8...v0.2.9
v0.2.8
What's Changed
- fix(dependencies): fix breaking change in FlagEmbedding 1.3 dependency by @rguilmont in #44
New Contributors
- @rguilmont made their first contribution in #44
Full Changelog: v0.2.7...v0.2.8
v0.2.7
What's Changed
Summary
- update Cohere's default rerank model to
rerank-english-v3.0
. - update Jina AI's default rerank model to
jina-reranker-v2-base-multilingual
. - update Voyage AI's default rerank model to
rerank-2
.
Full Changelog: v0.2.6...v0.2.7
v0.2.6
What's Changed
- fix(bm25): bump nltk to 3.9.1 & download punk_tab by @jean-humann in #37
- refactor: support jina embeddings v3 by @makram93 in #39
- feat: introduce a new param for v3 by @makram93 in #40
- fix: Update voyage, mistral, nomic embedding functions. by @wxywb in #41
New Contributors
- @jean-humann made their first contribution in #37
- @makram93 made their first contribution in #39
Summary
- Update JinaEmbeddingFunction to support v3.
- Update VoyageEmbeddingFunction to support latest models.
- Fix bugs in NomicEmbeddingFunction and MistralEmbeddingFunction.
Full Changelog: v0.2.5...v0.2.6
v0.2.5
v0.2.4
What's Changed
- feat: Add MistralEmbeddingFunction, NomicEmbeddingFunction and VertexAIEmbeddingFunction by @reina-w in #26
- feat: Add mGTE embedding function. by @wxywb in #27
- Update code to use csr_matrix and np.float64 for sparse vectors by @alesaccoia in #28
- feat: Add nomic and mistralai depencencies, remove the vertexai tempo… by @wxywb in #29
- Add instructor embedding and update Mistral embedding. by @wxywb in #30
- fix: Fix normalize embedding, Mistral's client API, changes BM25 default worker num to 1. by @wxywb in #31
Summarization
- Add MistralEmbeddingFunction, NomicEmbeddingFunction, MGTEEmbeddingFunction, InstructorEmbeddingFunction
- Fix 1d sparse embedding stacking. #25
- Fix SentenceTransformer normalization.
- Change BM25 default worker num from cpu_count() to 1.
New Contributors
- @reina-w made their first contribution in #26
- @alesaccoia made their first contribution in #28
Full Changelog: v0.2.3...v0.2.4
v0.2.3
v0.2.2
What's Changed
- Use onnx as default embedding function.
- Decouple model dependencies, installing them as needed.
Full Changelog: v0.2.1...v0.2.2