File size: 1,673 Bytes
a3c665b
 
 
 
 
 
 
 
 
 
bcb7072
a3c665b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
bcb7072
a3c665b
 
 
bcb7072
 
 
 
a3c665b
 
 
 
 
 
bcb7072
 
 
 
 
 
 
 
 
 
 
 
 
 
a3c665b
 
 
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
---
library_name: light-embed
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity

---

# LightEmbed/baai-llm-embedder-onnx

This is the ONNX version of the Sentence Transformers model BAAI/llm-embedder for sentence embedding, optimized for speed and lightweight performance. By utilizing onnxruntime and tokenizers instead of heavier libraries like sentence-transformers and transformers, this version ensures a smaller library size and faster execution. Below are the details of the model:
- Base model: BAAI/llm-embedder
- Embedding dimension: 768
- Max sequence length: 512
- File size on disk:  0.41 GB
- Pooling incorporated: Yes

This ONNX model consists all components in the original sentence transformer model:
Transformer, Pooling, Normalize

<!--- Describe your model here -->

## Usage (LightEmbed)

Using this model becomes easy when you have [LightEmbed](https://pypi.org/project/light-embed/) installed:

```
pip install -U light-embed
```

Then you can use the model using the original model name like this:

```python
from light_embed import TextEmbedding
sentences = [
	"This is an example sentence",
	"Each sentence is converted"
]

model = TextEmbedding('BAAI/llm-embedder')
embeddings = model.encode(sentences)
print(embeddings)
```

Then you can use the model using onnx model name like this:

```python
from light_embed import TextEmbedding
sentences = [
	"This is an example sentence",
	"Each sentence is converted"
]

model = TextEmbedding('LightEmbed/baai-llm-embedder-onnx')
embeddings = model.encode(sentences)
print(embeddings)
```

## Citing & Authors

Binh Nguyen / [email protected]