Create app.py

app.py

@@ -0,0 +1,53 @@
+import gradio as gr
+from transformers import AutoModelForSequenceClassification, AutoTokenizer, pipeline
+import torch
+import os
+
+# Path to your repository with renamed files
+model_path = "Ajay1311/bert-finetuned-phishing"
+
+# Define paths to renamed files
+weights_file = os.path.join(model_path, "phish_model.bin")
+config_file = os.path.join(model_path, "phish_config.json")
+tokenizer_path = model_path  # Directory containing tokenizer files
+
+# Load the tokenizer with renamed files
+tokenizer = AutoTokenizer.from_pretrained(
+    tokenizer_path,
+    config_file="phish_tokenizer_config.json",
+    vocab_file="phish_vocab.txt",
+    tokenizer_file="phih_tokenizer.json",
+    special_tokens_map_file="phish_special_tokens_map.json",
+    local_files_only=True
+)
+
+# Load the model with renamed files
+model = AutoModelForSequenceClassification.from_pretrained(
+    model_path,
+    config_file=config_file,
+    local_files_only=True
+)
+
+# Load the weights manually
+model.load_state_dict(torch.load(weights_file))
+
+# Create the pipeline for text classification
+classifier = pipeline("text-classification", model=model, tokenizer=tokenizer)
+
+# Define the prediction function
+def predict_phishing(text):
+    result = classifier(text)[0]
+    label = "Phishing" if result["label"] == "LABEL_1" else "Benign"
+    return f"Prediction: {label}\nConfidence: {result['score']:.4f}"
+
+# Create the Gradio interface
+iface = gr.Interface(
+    fn=predict_phishing,
+    inputs=gr.Textbox(lines=2, placeholder="Enter URL, SMS, or email text here..."),
+    outputs="text",
+    title="Phishing Detection with BERT",
+    description="Enter a URL, SMS, or email text to check if it's phishing or benign."
+)
+
+# Launch the interface
+iface.launch()