Update app.py

app.py

@@ -5,7 +5,8 @@ from langchain_community.embeddings import HuggingFaceEmbeddings
 from langchain_community.document_loaders import PyMuPDFLoader
 from langchain_text_splitters import CharacterTextSplitter
 from langchain.chains import RetrievalQA
-from transformers import pipeline  # Local model execution
+from langchain_community.llms import HuggingFacePipeline
+from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline
 
 def create_qa_system():
     try:
@@ -34,19 +35,27 @@ def create_qa_system():
         # Build vector store
         db = FAISS.from_documents(texts, embeddings)
         
-        # Local model pipeline
-        qa_pipeline = pipeline(
+        # Initialize local model with LangChain wrapper
+        model_name = "google/flan-t5-small"
+        tokenizer = AutoTokenizer.from_pretrained(model_name)
+        model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
+        
+        pipe = pipeline(
             "text2text-generation",
-            model="google/flan-t5-small",  # Runs locally
-            device=-1,  # Use CPU
+            model=model,
+            tokenizer=tokenizer,
             max_length=128,
-            temperature=0.2
+            temperature=0.2,
+            device_map="auto"
         )
         
+        llm = HuggingFacePipeline(pipeline=pipe)
+        
         return RetrievalQA.from_chain_type(
-            llm=qa_pipeline,
+            llm=llm,
             chain_type="stuff",
             retriever=db.as_retriever(search_kwargs={"k": 2}))
+        )
     except Exception as e:
         raise gr.Error(f"Initialization failed: {str(e)}")