Add application file

app.py

@@ -0,0 +1,74 @@
+import gradio as gr
+import torch
+from transformers import DistilBertTokenizer, DistilBertForSequenceClassification
+import torch.nn.functional as F
+
+# 1. Load the model and tokenizer from your saved directory
+# If uploading to Hugging Face Spaces, ensure your saved_model folder is uploaded too!
+model_path = "./saved_model" # Update this path to where your model is saved
+
+try:
+    tokenizer = DistilBertTokenizer.from_pretrained(model_path)
+    model = DistilBertForSequenceClassification.from_pretrained(model_path)
+    model.eval() # Set model to evaluation mode
+except Exception as e:
+    # Fallback to base model if you are just testing the code without your fine-tuned weights
+    # (Remove this try-except block in your final version, it's just to prevent crashes if path is wrong)
+    print("Could not load fine-tuned model, loading base model for demonstration purposes...")
+    tokenizer = DistilBertTokenizer.from_pretrained('distilbert-base-uncased')
+    model = DistilBertForSequenceClassification.from_pretrained('distilbert-base-uncased', num_labels=2)
+    model.eval()
+
+# 2. Define the prediction function
+def predict_spam(message):
+    if not message.strip():
+        return {"Please enter a message": 1.0}
+
+    # Tokenize input
+    inputs = tokenizer(message, return_tensors="pt", truncation=True, padding=True, max_length=128)
+    
+    # Get model prediction
+    with torch.no_grad():
+        outputs = model(**inputs)
+        logits = outputs.logits
+        
+    # Apply softmax to get probabilities
+    probabilities = F.softmax(logits, dim=1).squeeze()
+    
+    # Our labels are: 0 -> Ham, 1 -> Spam
+    prob_ham = probabilities[0].item()
+    prob_spam = probabilities[1].item()
+    
+    # Gradio expects a dictionary of {label: probability} for the Label component
+    return {"Ham (Legitimate)": prob_ham, "Spam (Malicious)": prob_spam}
+
+# 3. Create the Gradio Interface
+# We use a clean, modern interface
+demo = gr.Interface(
+    fn=predict_spam,
+    inputs=gr.Textbox(
+        lines=4, 
+        placeholder="Type an email or SMS message here...",
+        label="Message Content"
+    ),
+    outputs=gr.Label(num_top_classes=2, label="Prediction Confidence"),
+    title="🛡️ Spam Detection AI",
+    description="""
+    ### Deep Learning Project (2025)
+    This application uses a fine-tuned **DistilBERT** Transformer model to classify text messages as either Spam or Ham.
+    * Enter a message below and click Submit.
+    * **Examples of Spam:** 'URGENT! You have won a 1 week FREE membership. Call 087124006024'
+    * **Examples of Ham:** 'Hey, are we still meeting for lunch tomorrow?'
+    """,
+    examples=[
+        ["WINNER!! As a valued network customer you have been selected to receivea £900 prize reward! To claim call 09061701461. Claim code KL341. Valid 12 hours only."],
+        ["Hey man, just checking in. Are we still on for the movie tonight at 8?"],
+        ["URGENT: Your bank account has been locked due to suspicious activity. Please click this link to verify your identity: http://secure-login-update.com"],
+        ["I'll be about 10 minutes late, stuck in traffic."]
+    ],
+    theme=gr.themes.Soft()
+)
+
+# 4. Launch the app
+if __name__ == "__main__":
+    demo.launch()

requirements.txt

@@ -0,0 +1,3 @@
+gradio
+transformers
+torch

saved_model/config.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:faf37237976cc13bad4520548a113c77e7529c3b4ff6b598648fbce1696499e3
+size 663

saved_model/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c07d77d4bad466798bc928c36a02d91cadbef0c2047def7e445928f1825d959
+size 267832560

saved_model/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0b616fb2661c2ad4dbc6c0ee5c4c1caefd532989e8c399005744f5442e4116ed
+size 535724875

saved_model/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c0f6bef14c378cf6dadcce9566f94c89a1de5a87f6dab300dc2cfb3620974e5
+size 14455

saved_model/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0fd612b91f9c87518d4d69e324f23842dc16c0f0824a97f416850db8d65d6175
+size 1465

saved_model/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d241a60d5e8f04cc1b2b3e9ef7a4921b27bf526d9f6050ab90f9267a1f9e5c66
+size 711396

saved_model/tokenizer_config.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:af5ecd1e2aa3d659df1ea773292e80eb41a5c3b4026126019e5c6dcdb7cbc044
+size 328

saved_model/trainer_state.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4748a47ae13b6f9fd0182a6f65f4ee77883b846e370684a964475170d957f80
+size 5955

saved_model/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ab8b54e46b16083bb42102640d659199bfe96644c6fe4869e1135f8028f75f5c
+size 5201