File size: 1,446 Bytes
0ded642
 
920e680
0ded642
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: apache-2.0
pipeline_tag: image-classification
---
# About
This is a fork of MichalMlodawski/nsfw-image-detection-large which became unavailable.

# Usage example
```
from PIL import Image
import torch
from transformers import AutoProcessor, FocalNetForImageClassification

DEVICE = torch.device("cuda")
model_path = "lovetillion/nsfw-image-detection-large"

# Load the model and feature extractor
feature_extractor = AutoProcessor.from_pretrained(model_path)
model = FocalNetForImageClassification.from_pretrained(model_path).to(DEVICE)
model.eval()

# Mapping from model labels to NSFW categories
label_to_category = {
    "LABEL_0": "Safe",
    "LABEL_1": "Questionable",
    "LABEL_2": "Unsafe"
}

filename = "example.png"
image = Image.open(filename)
inputs = feature_extractor(images=image, return_tensors="pt")
inputs.to(DEVICE)

with torch.no_grad():
    outputs = model(**inputs)
    probabilities = torch.nn.functional.softmax(outputs.logits, dim=-1)
    confidence, predicted = torch.max(probabilities, 1)
label = model.config.id2label[predicted.item()]

if label != "SAFE":
    print( label, confidence.item() * 100, filename )
else:
    print( label, confidence.item() * 100, filename )
```

# For more information

* Live demonstration in a production ensemble workflow: https://piglet.video
* Results from our ethical AI whitepaper: https://lovetillion.org/liaise.pdf
* Join us on Telegram at https://t.me/pigletproject