fastembed 5.16.1

Library for generating vector embeddings, reranking locally.
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
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143

use std::{fmt::Display, str::FromStr};

use super::model_info::ModelInfo;

#[derive(Debug, Default, Clone, PartialEq, Eq)]
pub enum ImageEmbeddingModel {
    /// Qdrant/clip-ViT-B-32-vision
    #[default]
    ClipVitB32,
    /// Qdrant/resnet50-onnx
    Resnet50,
    /// Qdrant/Unicom-ViT-B-16
    UnicomVitB16,
    /// Qdrant/Unicom-ViT-B-32
    UnicomVitB32,
    /// nomic-ai/nomic-embed-vision-v1.5
    NomicEmbedVisionV15,
}

pub fn models_list() -> Vec<ModelInfo<ImageEmbeddingModel>> {
    let models_list = vec![
        ModelInfo {
            model: ImageEmbeddingModel::ClipVitB32,
            dim: 512,
            description: String::from("CLIP vision encoder based on ViT-B/32"),
            model_code: String::from("Qdrant/clip-ViT-B-32-vision"),
            model_file: String::from("model.onnx"),
            additional_files: Vec::new(),
            output_key: None,
        },
        ModelInfo {
            model: ImageEmbeddingModel::Resnet50,
            dim: 2048,
            description: String::from("ResNet-50 from `Deep Residual Learning for Image Recognition <https://arxiv.org/abs/1512.03385>`__."),
            model_code: String::from("Qdrant/resnet50-onnx"),
            model_file: String::from("model.onnx"),
            additional_files: Vec::new(),
            output_key: None,
        },
        ModelInfo {
            model: ImageEmbeddingModel::UnicomVitB16,
            dim: 768,
            description: String::from("Unicom Unicom-ViT-B-16 from open-metric-learning"),
            model_code: String::from("Qdrant/Unicom-ViT-B-16"),
            model_file: String::from("model.onnx"),
            additional_files: Vec::new(),
            output_key: None,
        },
        ModelInfo {
            model: ImageEmbeddingModel::UnicomVitB32,
            dim: 512,
            description: String::from("Unicom Unicom-ViT-B-32 from open-metric-learning"),
            model_code: String::from("Qdrant/Unicom-ViT-B-32"),
            model_file: String::from("model.onnx"),
            additional_files: Vec::new(),
            output_key: None,
        },
        ModelInfo {
            model: ImageEmbeddingModel::NomicEmbedVisionV15,
            dim: 768,
            description: String::from("Nomic NomicEmbedVisionV15"),
            model_code: String::from("nomic-ai/nomic-embed-vision-v1.5"),
            model_file: String::from("onnx/model.onnx"),
            additional_files: Vec::new(),
            output_key: None,
        },
    ];

    // TODO: Use when out in stable
    // assert_eq!(
    //     std::mem::variant_count::<ImageEmbeddingModel>(),
    //     models_list.len(),
    //     "models::models() is not exhaustive"
    // );

    models_list
}

impl Display for ImageEmbeddingModel {
    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
        let model_info = models_list()
            .into_iter()
            .find(|model| model.model == *self)
            .ok_or(std::fmt::Error)?;
        write!(f, "{}", model_info.model_code)
    }
}

impl FromStr for ImageEmbeddingModel {
    type Err = String;

    fn from_str(s: &str) -> Result<Self, Self::Err> {
        models_list()
            .into_iter()
            .find(|m| m.model_code.eq_ignore_ascii_case(s))
            .map(|m| m.model)
            .ok_or_else(|| format!("Unknown embedding model: {s}"))
    }
}

impl TryFrom<String> for ImageEmbeddingModel {
    type Error = String;

    fn try_from(value: String) -> Result<Self, Self::Error> {
        value.parse()
    }
}

#[cfg(test)]
pub(crate) fn all_variants() -> Vec<ImageEmbeddingModel> {
    fn _exhaustive_guard(m: &ImageEmbeddingModel) {
        match m {
            ImageEmbeddingModel::ClipVitB32 => (),
            ImageEmbeddingModel::Resnet50 => (),
            ImageEmbeddingModel::UnicomVitB16 => (),
            ImageEmbeddingModel::UnicomVitB32 => (),
            ImageEmbeddingModel::NomicEmbedVisionV15 => (),
        }
    }
    vec![
        ImageEmbeddingModel::ClipVitB32,
        ImageEmbeddingModel::Resnet50,
        ImageEmbeddingModel::UnicomVitB16,
        ImageEmbeddingModel::UnicomVitB32,
        ImageEmbeddingModel::NomicEmbedVisionV15,
    ]
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn every_variant_has_model_info() {
        let listed: Vec<_> = models_list().into_iter().map(|i| i.model).collect();
        for variant in all_variants() {
            assert!(
                listed.contains(&variant),
                "{variant:?} is missing from models_list(); get_model_info would panic"
            );
        }
    }
}