Add Transformers.js and WebNN example to README.md
Browse files
README.md
CHANGED
|
@@ -10,6 +10,7 @@ license: apache-2.0
|
|
| 10 |
pipeline_tag: text-classification
|
| 11 |
tags:
|
| 12 |
- Intel
|
|
|
|
| 13 |
model-index:
|
| 14 |
- name: polite-guard
|
| 15 |
results:
|
|
@@ -20,10 +21,10 @@ model-index:
|
|
| 20 |
type: polite-guard
|
| 21 |
metrics:
|
| 22 |
- type: accuracy
|
| 23 |
-
value: 92
|
| 24 |
name: Accuracy
|
| 25 |
- type: f1
|
| 26 |
-
value: 92
|
| 27 |
name: F1 Score
|
| 28 |
---
|
| 29 |
# Polite Guard
|
|
@@ -82,8 +83,8 @@ The code for the synthetic data generation and fine-tuning can be found [here](h
|
|
| 82 |
|
| 83 |
Here are the key performance metrics of the model on the test dataset containing both synthetic and manually annotated data:
|
| 84 |
|
| 85 |
-
- **Accuracy**: 92
|
| 86 |
-
- **F1-Score**: 92
|
| 87 |
|
| 88 |
## How to Use
|
| 89 |
|
|
@@ -92,9 +93,33 @@ You can use this model directly with a pipeline for categorizing text into class
|
|
| 92 |
```python
|
| 93 |
from transformers import pipeline
|
| 94 |
|
| 95 |
-
classifier = pipeline("text-classification",
|
| 96 |
text = "Your input text"
|
| 97 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 98 |
```
|
| 99 |
## Articles
|
| 100 |
|
|
|
|
| 10 |
pipeline_tag: text-classification
|
| 11 |
tags:
|
| 12 |
- Intel
|
| 13 |
+
- transformers.js
|
| 14 |
model-index:
|
| 15 |
- name: polite-guard
|
| 16 |
results:
|
|
|
|
| 21 |
type: polite-guard
|
| 22 |
metrics:
|
| 23 |
- type: accuracy
|
| 24 |
+
value: 92
|
| 25 |
name: Accuracy
|
| 26 |
- type: f1
|
| 27 |
+
value: 92
|
| 28 |
name: F1 Score
|
| 29 |
---
|
| 30 |
# Polite Guard
|
|
|
|
| 83 |
|
| 84 |
Here are the key performance metrics of the model on the test dataset containing both synthetic and manually annotated data:
|
| 85 |
|
| 86 |
+
- **Accuracy**: 92% on the Polite Guard test dataset.
|
| 87 |
+
- **F1-Score**: 92% on the Polite Guard test dataset.
|
| 88 |
|
| 89 |
## How to Use
|
| 90 |
|
|
|
|
| 93 |
```python
|
| 94 |
from transformers import pipeline
|
| 95 |
|
| 96 |
+
classifier = pipeline("text-classification", "Intel/polite-guard")
|
| 97 |
text = "Your input text"
|
| 98 |
+
output = classifier(text)
|
| 99 |
+
print(output)
|
| 100 |
+
```
|
| 101 |
+
|
| 102 |
+
The next example demonstrates how to run this model in the browser using Hugging Face's `transformers.js` library with `webnn-gpu` for hardware acceleration.
|
| 103 |
+
|
| 104 |
+
```html
|
| 105 |
+
<!DOCTYPE html>
|
| 106 |
+
<html>
|
| 107 |
+
<body>
|
| 108 |
+
<h1>WebNN Transformers.js Intel/polite-guard</h1>
|
| 109 |
+
<script type="module">
|
| 110 |
+
import { pipeline } from "https://cdn.jsdelivr.net/npm/@huggingface/transformers";
|
| 111 |
+
|
| 112 |
+
const classifier = await pipeline("text-classification", "Intel/polite-guard", {
|
| 113 |
+
dtype: "fp32",
|
| 114 |
+
device: "webnn-gpu", // You can also try: "webgpu", "webnn", "webnn-npu", "webnn-cpu", "wasm"
|
| 115 |
+
});
|
| 116 |
+
|
| 117 |
+
const text = "Your input text";
|
| 118 |
+
const output = await classifier(text);
|
| 119 |
+
console.log(`${text}: ${output[0].label}`);
|
| 120 |
+
</script>
|
| 121 |
+
</body>
|
| 122 |
+
</html>
|
| 123 |
```
|
| 124 |
## Articles
|
| 125 |
|