Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import torch
|
2 |
+
import gradio as gr
|
3 |
+
import numpy as np
|
4 |
+
import matplotlib.pyplot as plt
|
5 |
+
from transformers import AutoTokenizer, AutoModelForSequenceClassification
|
6 |
+
|
7 |
+
torch.set_num_threads(torch.get_num_threads())
|
8 |
+
|
9 |
+
# Load the trained model and tokenizer from Hugging Face Hub
|
10 |
+
model_path = "HyperX-Sentience/RogueBERT-Toxicity-85K"
|
11 |
+
model = AutoModelForSequenceClassification.from_pretrained(model_path)
|
12 |
+
tokenizer = AutoTokenizer.from_pretrained(model_path)
|
13 |
+
|
14 |
+
# Move the model to CUDA if available
|
15 |
+
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
16 |
+
model.to(device)
|
17 |
+
|
18 |
+
# Define toxicity labels
|
19 |
+
labels = ["toxic", "severe_toxic", "obscene", "threat", "insult", "identity_hate"]
|
20 |
+
|
21 |
+
def predict_toxicity(comment):
|
22 |
+
"""Predicts the toxicity levels of a given comment."""
|
23 |
+
inputs = tokenizer(comment, truncation=True, padding="max_length", max_length=128, return_tensors="pt")
|
24 |
+
inputs = {key: val.to(device) for key, val in inputs.items()}
|
25 |
+
|
26 |
+
with torch.no_grad():
|
27 |
+
outputs = model(**inputs)
|
28 |
+
probabilities = torch.sigmoid(outputs.logits).cpu().numpy()[0]
|
29 |
+
|
30 |
+
return {labels[i]: float(probabilities[i]) for i in range(len(labels))}
|
31 |
+
|
32 |
+
def visualize_toxicity(comment):
|
33 |
+
"""Generates a bar chart showing toxicity levels."""
|
34 |
+
scores = predict_toxicity(comment)
|
35 |
+
|
36 |
+
# Create bar chart
|
37 |
+
plt.figure(figsize=(6, 4))
|
38 |
+
plt.bar(scores.keys(), scores.values(), color=['blue', 'red', 'green', 'purple', 'orange', 'brown'])
|
39 |
+
plt.ylim(0, 1)
|
40 |
+
plt.ylabel("Toxicity Score")
|
41 |
+
plt.title("Toxicity Analysis")
|
42 |
+
plt.xticks(rotation=45)
|
43 |
+
plt.grid(axis='y', linestyle='--', alpha=0.7)
|
44 |
+
|
45 |
+
# Save plot to display in Gradio
|
46 |
+
plt.savefig("toxicity_plot.png")
|
47 |
+
plt.close()
|
48 |
+
|
49 |
+
return "toxicity_plot.png"
|
50 |
+
|
51 |
+
# Gradio interface
|
52 |
+
demo = gr.Interface(
|
53 |
+
fn=visualize_toxicity,
|
54 |
+
inputs=gr.Textbox(label="Enter a comment:"),
|
55 |
+
outputs=gr.Image(type="file", label="Toxicity Scores"),
|
56 |
+
title="Toxicity Detection with RogueBERT",
|
57 |
+
description="Enter a comment to analyze its toxicity levels. The results will be displayed as a bar chart."
|
58 |
+
)
|
59 |
+
|
60 |
+
demo.launch()
|