Upload 4 files

app (3).py

@@ -0,0 +1,59 @@
+# -*- coding: utf-8 -*-
+"""app.ipynb
+
+Automatically generated by Colab.
+
+Original file is located at
+    https://colab.research.google.com/drive/1P96bh2DkRkgFJiX-Vvr3WeI6gsfDMStd
+"""
+
+import tensorflow as tf
+import gradio as gr
+import numpy as np
+from tensorflow.keras.preprocessing.text import Tokenizer
+from tensorflow.keras.preprocessing.sequence import pad_sequences
+
+# 1. Load the trained model
+model = tf.keras.models.load_model("spam_detection_model.h5")
+
+# 2. Rebuild the tokenizer inside code (no external file needed)
+# These are sample training-like sentences used to refit the tokenizer
+sample_messages = [
+    "Congratulations! You've won a prize.",
+    "Free entry in a contest!",
+    "Call now to claim your reward!",
+    "URGENT: Your number was selected!",
+    "Meeting at 10 AM tomorrow.",
+    "Lunch at 1 PM?",
+    "Please submit the report by tonight.",
+    "Let's catch up for the project discussion.",
+    "Click here for a special discount!",
+    "This is not spam, just a reminder.",
+]
+
+# 3. Tokenizer setup (same as used during training)
+tokenizer = Tokenizer(num_words=5000, oov_token="<OOV>")
+tokenizer.fit_on_texts(sample_messages)
+
+# 4. Set max sequence length
+max_length = 100
+
+# 5. Prediction function
+def predict_spam(message):
+    seq = tokenizer.texts_to_sequences([message])
+    padded = pad_sequences(seq, maxlen=max_length, padding='post')
+    pred = float(model.predict(padded, verbose=0)[0])
+    label = "🔴 Spam" if pred > 0.5 else "🟢 Ham"
+    return f"{label} ({pred * 100:.2f}% confidence)"
+
+# 6. Gradio Interface
+iface = gr.Interface(
+    fn=predict_spam,
+    inputs=gr.Textbox(lines=3, placeholder="Enter a message to classify..."),
+    outputs="text",
+    title="📩 Spam or Ham Classifier",
+    description="This LSTM-based app classifies text messages as Spam or Ham. Built with TensorFlow and Gradio.",
+    theme="default"
+)
+
+iface.launch()

requirements.txt

@@ -0,0 +1,5 @@
+pandas
+numpy
+tensorflow
+scikit-learn
+gradio

spam_detection_model.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e41ea0703d53b157c12ff1dcc8fb9188a197b32f4b224b3d800f6b92f2371da9
+size 4301280