Create app.ppy

app.ppy

@@ -0,0 +1,52 @@
+import gradio as gr
+from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
+from sentence_transformers import SentenceTransformer
+from sklearn.metrics.pairwise import cosine_similarity
+import numpy as np
+
+# FAQ Data
+faq_data = [
+    {"question": "What is AI?", "answer": "AI stands for Artificial Intelligence."},
+    {"question": "What is IBM Granite?", "answer": "Granite is IBM’s family of foundation models."},
+    {"question": "What is Streamlit?", "answer": "Streamlit is a Python library to build interactive web apps."},
+    {"question": "What is machine learning?", "answer": "Machine learning is a subset of AI involving data-driven models."}
+]
+
+# Load models
+embedder = SentenceTransformer('all-MiniLM-L6-v2')
+tokenizer = AutoTokenizer.from_pretrained("ibm/granite-13b-instruct")
+model = AutoModelForCausalLM.from_pretrained("ibm/granite-13b-instruct", torch_dtype="auto", device_map="auto")
+qa_pipeline = pipeline("text-generation", model=model, tokenizer=tokenizer)
+
+# Semantic Search
+def find_most_similar_question(user_question, faq_data):
+    user_vec = embedder.encode([user_question])
+    faq_vecs = embedder.encode([faq["question"] for faq in faq_data])
+    similarities = cosine_similarity(user_vec, faq_vecs)[0]
+    best_idx = int(np.argmax(similarities))
+    return faq_data[best_idx]["question"], faq_data[best_idx]["answer"]
+
+# Chatbot logic
+def chatbot(user_question):
+    matched_q, matched_ans = find_most_similar_question(user_question, faq_data)
+
+    prompt = f"""You are a helpful assistant. A user asked: "{user_question}"
+Here is the most relevant question and answer from FAQ:
+Q: {matched_q}
+A: {matched_ans}
+
+Now generate a helpful and clear answer based on the above."""
+
+    output = qa_pipeline(prompt, max_new_tokens=150, do_sample=False)[0]["generated_text"]
+    final_response = output[len(prompt):].strip()
+    return final_response
+
+# Gradio UI
+iface = gr.Interface(fn=chatbot, 
+                     inputs=gr.Textbox(lines=2, placeholder="Ask a question..."), 
+                     outputs="text",
+                     title="🤖 FAQ Chatbot with IBM Granite",
+                     description="Ask any question related to the FAQ and get an AI-generated answer using IBM Granite.")
+
+if __name__ == "__main__":
+    iface.launch()