Xenova HF staff commited on
Commit
77722e7
1 Parent(s): 7ca943b

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +48 -0
README.md CHANGED
@@ -4,4 +4,52 @@ library_name: "transformers.js"
4
 
5
  https://huggingface.co/thenlper/gte-small with ONNX weights to be compatible with Transformers.js.
6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7
  Note: Having a separate repo for ONNX weights is intended to be a temporary solution until WebML gains more traction. If you would like to make your models web-ready, we recommend converting to ONNX using [🤗 Optimum](https://huggingface.co/docs/optimum/index) and structuring your repo like this one (with ONNX weights located in a subfolder named `onnx`).
 
4
 
5
  https://huggingface.co/thenlper/gte-small with ONNX weights to be compatible with Transformers.js.
6
 
7
+ ## Usage (Transformers.js)
8
+
9
+ If you haven't already, you can install the [Transformers.js](https://huggingface.co/docs/transformers.js) JavaScript library from [NPM](https://www.npmjs.com/package/@xenova/transformers) using:
10
+ ```bash
11
+ npm i @xenova/transformers
12
+ ```
13
+
14
+ You can then use the model to compute embeddings like this:
15
+
16
+ ```js
17
+ import { pipeline } from '@xenova/transformers';
18
+
19
+ // Create a feature-extraction pipeline
20
+ const extractor = await pipeline('feature-extraction', 'Xenova/gte-small');
21
+
22
+ // Compute sentence embeddings
23
+ const sentences = ['That is a happy person', 'That is a very happy person'];
24
+ const output = await extractor(sentences, { pooling: 'mean', normalize: true });
25
+ console.log(output);
26
+ // Tensor {
27
+ // dims: [ 2, 384 ],
28
+ // type: 'float32',
29
+ // data: Float32Array(768) [ -0.053555335849523544, 0.00843878649175167, ... ],
30
+ // size: 768
31
+ // }
32
+
33
+ // Compute cosine similarity
34
+ import { cos_sim } from '@xenova/transformers';
35
+ console.log(cos_sim(output[0].data, output[1].data))
36
+ // 0.9798319649182318
37
+ ```
38
+
39
+ You can convert this Tensor to a nested JavaScript array using `.tolist()`:
40
+ ```js
41
+ console.log(output.tolist());
42
+ // [
43
+ // [ -0.053555335849523544, 0.00843878649175167, 0.06234041228890419, ... ],
44
+ // [ -0.049980051815509796, 0.03879701718688011, 0.07510733604431152, ... ]
45
+ // ]
46
+ ```
47
+
48
+ By default, an 8-bit quantized version of the model is used, but you can choose to use the full-precision (fp32) version by specifying `{ quantized: false }` in the `pipeline` function:
49
+ ```js
50
+ const extractor = await pipeline('feature-extraction', 'Xenova/gte-small', { quantized: false });
51
+ ```
52
+
53
+ ---
54
+
55
  Note: Having a separate repo for ONNX weights is intended to be a temporary solution until WebML gains more traction. If you would like to make your models web-ready, we recommend converting to ONNX using [🤗 Optimum](https://huggingface.co/docs/optimum/index) and structuring your repo like this one (with ONNX weights located in a subfolder named `onnx`).