Update app.py

app.py

@@ -1,29 +1,68 @@
 import gradio as gr
 from transformers import pipeline
 
-# Use a smaller model that fits in 16GB RAM without quantization
-model = pipeline(
-    "text-generation",
-    model="HuggingFaceH4/zephyr-7b-beta",
-    device_map="auto",
-    torch_dtype="auto"  # Automatically optimizes precision
-)
-
-def respond(prompt):
-    try:
-        response = model(
-            f"<|user|>{prompt}</s><|assistant|>",
-            max_new_tokens=40,  # Shorter responses save memory
-            do_sample=True,
-            temperature=0.7
-        )
-        return response[0]['generated_text'].split("<|assistant|>")[-1].strip()
-    except Exception as e:
-        return f"Error: {str(e)}"
-
-gr.Interface(
-    fn=respond,
-    inputs=gr.Textbox(placeholder="Ask me anything..."),
-    outputs="text",
-    title="🤖 Optimized Chatbot"
-).launch()
+# --- 1. LOAD OUR AI MODELS ---
+# We load the models once when the app starts.
+# This is more efficient than loading them for every request.
+question_answerer = pipeline("question-answering", model="distilbert-base-uncased-distilled-squad")
+summarizer = pipeline("summarization", model="sshleifer/distilbart-cnn-6-6")
+
+
+# --- 2. DEFINE THE CORE FUNCTIONS ---
+# These functions will take user input and use the appropriate model.
+
+def answer_question(context, question):
+    """Uses the QA model to find an answer within a given text."""
+    if not context or not question:
+        return "(Please provide both context and a question.)"
+    result = question_answerer(question=question, context=context)
+    return result['answer']
+
+def summarize_text(text):
+    """Uses the Summarization model to shorten a piece of text."""
+    if not text:
+        return "(Please provide text to summarize.)"
+    # We add some parameters for better, shorter summaries on a CPU
+    summary = summarizer(text, max_length=50, min_length=25, do_sample=False)
+    return summary[0]['summary_text']
+
+
+# --- 3. BUILD THE GRADIO INTERFACE ---
+# This is where we design the web app's layout and connect it to our functions.
+
+with gr.Blocks() as demo:
+    gr.Markdown("# Ultimate AI Assistant (CPU Edition) 🤖")
+    gr.Markdown("An experiment in combining multiple small, efficient AI models into one application.")
+
+    with gr.Tabs():
+        # --- First Tab: Question Answering ---
+        with gr.TabItem("❓ Ask a Question"):
+            gr.Markdown("Give the AI some text (context) and ask a question about it.")
+            with gr.Row():
+                qa_context_input = gr.Textbox(lines=7, label="Context", placeholder="Paste a paragraph or article here...")
+                qa_question_input = gr.Textbox(label="Question", placeholder="What do you want to know?")
+            qa_button = gr.Button("Get Answer")
+            qa_output = gr.Textbox(label="Answer")
+
+        # --- Second Tab: Summarization ---
+        with gr.TabItem("📚 Summarize Text"):
+            gr.Markdown("Paste in a long piece of text and the AI will create a short summary.")
+            summarize_input = gr.Textbox(lines=10, label="Text to Summarize", placeholder="Paste a long article or text here...")
+            summarize_button = gr.Button("Summarize")
+            summarize_output = gr.Textbox(label="Summary")
+
+    # --- 4. Connect Buttons to Functions ---
+    qa_button.click(
+        fn=answer_question,
+        inputs=[qa_context_input, qa_question_input],
+        outputs=qa_output
+    )
+
+    summarize_button.click(
+        fn=summarize_text,
+        inputs=summarize_input,
+        outputs=summarize_output
+    )
+
+# --- 5. LAUNCH THE APP! ---
+demo.launch()