Spaces:
Running
Running
File size: 1,748 Bytes
43d7102 9e6277b 43d7102 |
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 |
# Load model directly
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import numpy as np
import torch
# Check if CUDA is available
if torch.cuda.is_available():
# Choose a specific GPU or use the default
device = torch.device("cuda:0")
else:
# Or CPU
device = torch.device("cpu")
tokenizer = AutoTokenizer.from_pretrained("kmack/malicious-url-detection")
model = AutoModelForSequenceClassification.from_pretrained("kmack/malicious-url-detection")
# set Model to cude
model = model.to(device)
# predict function
def get_predit(input_text: str) -> dict:
label2id = model.config.label2id
inputs = tokenizer(input_text, return_tensors='pt', truncation=True)
inputs = inputs.to(device)
outputs = model(**inputs)
logits = outputs.logits
sigmoid = torch.nn.Sigmoid()
probs = sigmoid(logits.squeeze().cpu())
probs = probs.detach().numpy()
for i, k in enumerate(label2id.keys()):
label2id[k] = probs[i]
label2id = {k: float(v) for k, v in sorted(label2id.items(), key=lambda item: item[1].item(), reverse=True)}
return label2id
# Define example URLs
example_url_1 = 'https://medium.com'
example_url_2 = 'http://google.com-redirect@valimail.com'
example_url_3 = 'https://a101-nisan-kampanyalari.com'
# Create the Gradio interface
demo = gr.Interface(
fn=get_predit,
inputs=gr.components.Textbox(label='Input', placeholder='Enter URL here...'),
outputs=gr.components.Label(label='Predictions', num_top_classes=5),
title='kmack/malicious-url-detection',
description='Detects whether a given URL is benign or potentially malicious.',
examples=[[example_url_1], [example_url_2], [example_url_3]],
allow_flagging='never'
)
demo.launch()
|