Update README.md
Browse files
README.md
CHANGED
@@ -2,6 +2,59 @@
|
|
2 |
library_name: "transformers.js"
|
3 |
---
|
4 |
|
5 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
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`).
|
|
|
2 |
library_name: "transformers.js"
|
3 |
---
|
4 |
|
5 |
+
|
6 |
+
# Whisper
|
7 |
+
|
8 |
+
[openai/whisper-tiny.en](https://huggingface.co/openai/whisper-tiny.en) with ONNX weights to be compatible with [Transformers.js](https://huggingface.co/docs/transformers.js).
|
9 |
+
|
10 |
+
|
11 |
+
## Usage
|
12 |
+
|
13 |
+
**Example:** Transcribe English.
|
14 |
+
|
15 |
+
|
16 |
+
```js
|
17 |
+
let url = 'https://huggingface.co/datasets/Xenova/transformers.js-docs/resolve/main/jfk.wav';
|
18 |
+
let transcriber = await pipeline('automatic-speech-recognition', 'Xenova/whisper-tiny.en');
|
19 |
+
let output = await transcriber(url);
|
20 |
+
// { text: " And so my fellow Americans ask not what your country can do for you, ask what you can do for your country." }
|
21 |
+
```
|
22 |
+
|
23 |
+
**Example:** Transcribe English w/ timestamps.
|
24 |
+
|
25 |
+
```js
|
26 |
+
let url = 'https://huggingface.co/datasets/Xenova/transformers.js-docs/resolve/main/jfk.wav';
|
27 |
+
let transcriber = await pipeline('automatic-speech-recognition', 'Xenova/whisper-tiny.en');
|
28 |
+
let output = await transcriber(url, { return_timestamps: true });
|
29 |
+
// {
|
30 |
+
// text: " And so my fellow Americans ask not what your country can do for you, ask what you can do for your country."
|
31 |
+
// chunks: [
|
32 |
+
// { timestamp: [0, 8], text: " And so my fellow Americans ask not what your country can do for you" }
|
33 |
+
// { timestamp: [8, 11], text: " ask what you can do for your country." }
|
34 |
+
// ]
|
35 |
+
// }
|
36 |
+
```
|
37 |
+
|
38 |
+
**Example:** Transcribe English w/ word-level timestamps.
|
39 |
+
|
40 |
+
```js
|
41 |
+
let url = 'https://huggingface.co/datasets/Xenova/transformers.js-docs/resolve/main/jfk.wav';
|
42 |
+
let transcriber = await pipeline('automatic-speech-recognition', 'Xenova/whisper-tiny.en');
|
43 |
+
let output = await transcriber(url, { return_timestamps: 'word' });
|
44 |
+
// {
|
45 |
+
// "text": " And so my fellow Americans ask not what your country can do for you ask what you can do for your country.",
|
46 |
+
// "chunks": [
|
47 |
+
// { "text": " And", "timestamp": [0, 0.78] },
|
48 |
+
// { "text": " so", "timestamp": [0.78, 1.06] },
|
49 |
+
// { "text": " my", "timestamp": [1.06, 1.46] },
|
50 |
+
// ...
|
51 |
+
// { "text": " for", "timestamp": [9.72, 9.92] },
|
52 |
+
// { "text": " your", "timestamp": [9.92, 10.22] },
|
53 |
+
// { "text": " country.", "timestamp": [10.22, 13.5] }
|
54 |
+
// ]
|
55 |
+
// }
|
56 |
+
```
|
57 |
+
|
58 |
+
---
|
59 |
|
60 |
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`).
|