candding 0.1.0

Pure candle embeddings for Rust with no ONNX: dense, sparse, multi-vector and rerank models today, image models on the roadmap
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73

#![doc = include_str!("docs.md")]

pub mod device;
#[doc(hidden)]
pub mod embed;
#[doc(hidden)]
pub mod error;
#[doc(hidden)]
pub mod hub;
#[doc(hidden)]
pub mod layers;
#[doc(hidden)]
pub mod loader;
#[doc(hidden)]
pub mod models;
pub mod multi_vector;
#[doc(hidden)]
pub mod pooling;
pub mod registry;
#[doc(hidden)]
pub mod rerank;
#[doc(hidden)]
pub mod sparse;
#[doc(hidden)]
pub mod template;
#[doc(hidden)]
pub mod tokenizer;
#[doc(hidden)]
pub mod weights;

pub use candle_core::DType;
pub use candle_core::Device;
pub use embed::Bm25TextEmbedding;
pub use embed::Bm25TextEmbeddingBuilder;
pub use embed::MultiVectorTextEmbedding;
pub use embed::MultiVectorTextEmbeddingBuilder;
pub use embed::SparseTextEmbedding;
pub use embed::SparseTextEmbeddingBuilder;
pub use embed::TextCrossEncoder;
pub use embed::TextCrossEncoderBuilder;
pub use embed::TextEmbedding;
pub use embed::TextEmbeddingBuilder;
pub use error::CanddingError;
pub use error::LengthKind;
pub use error::Result;
pub use models::MultiVectorHead;
pub use models::RerankHead;
pub use models::SparseHead;
pub use models::Unembedding;
pub use models::jina_bert::AlibiSlopeVariant;
pub use models::xlm_roberta_flash::PositionScheme;
pub use multi_vector::MaskedRowFill;
pub use multi_vector::MultiVectorEmbedding;
pub use multi_vector::ScoreReduction;
pub use pooling::Pooling;
pub use registry::Backend;
pub use registry::BackendSupport;
pub use registry::Catalog;
pub use registry::DtypeBackendSupport;
pub use registry::Family;
pub use registry::ModelDescriptor;
pub use registry::Support;
pub use registry::WeightsFormat;
pub use registry::multi_vector::MultiVectorDescriptor;
pub use registry::rerank::RerankDescriptor;
pub use registry::rerank::RerankInput;
pub use registry::sparse::SparseComputation;
pub use registry::sparse::SparseDescriptor;
pub use rerank::RerankScore;
pub use rerank::ScoreActivation;
pub use sparse::SparseEmbedding;
pub use tokenizer::PairTruncationPolicy;
pub use tokenizer::PromptTemplate;