updated app.py

app.py

@@ -81,9 +81,16 @@ def process_files(files):
     result = collection.get(include=['embeddings', 'documents', 'metadatas'])
 
     # HF Pipeline
-    hf_pipeline = pipeline("document-question-answering", model="impira/layoutlm-document-qa")
+    hf_pipeline = pipeline(
+    "text-generation",
+    model="mistralai/Mistral-7B-Instruct-v0.1",
+    torch_dtype="auto",
+    device_map="auto",  # Uses GPU if available
+    max_new_tokens=256,  # Adjust output length
+    )
 
-    llm = ChatHuggingFace(model_name="HuggingFaceH4/zephyr-7b-alpha")
+    # Wrap it in LangChain
+    llm = HuggingFacePipeline(pipeline=hf_pipeline)
     memory = ConversationBufferMemory(memory_key='chat_history', return_messages=True)
     retriever = vectorstore.as_retriever(search_kwargs={"k": 10})
     global conversation_chain