Update space

app.py

@@ -7,7 +7,7 @@ from langchain_ollama.llms import OllamaLLM
 
 from langchain_community.llms.huggingface_pipeline import HuggingFacePipeline
 
-from langchain import hub
+#from langchain import hub
 from langchain_core.runnables import RunnableParallel
 from langchain_core.runnables import RunnablePassthrough
 from langchain_core.output_parsers import StrOutputParser
@@ -20,9 +20,9 @@ For more information on `huggingface_hub` Inference API support, please check th
 
 # LLM Model#
 llm = HuggingFacePipeline.from_model_id(
-    model_id="meta-llama/Llama-3.2-3B",
+    model_id="llmware/bling-phi-3-gguf",
     task="text-generation",
-    pipeline_kwargs={"max_new_tokens": 10},
+    pipeline_kwargs={"max_new_tokens": 100},
 )
 
 # Initialize embedding model "all-MiniLM-L6-v2"
@@ -33,7 +33,10 @@ vector_store = Chroma(persist_directory="./chroma_db", embedding_function=embedd
 
 # See full prompt at https://smith.langchain.com/hub/rlm/rag-prompt
 # Basically a solid prompt for RAG 
-prompt = hub.pull("rlm/rag-prompt")
+prompt = """You are an assistant for question-answering tasks. Use the following pieces of retrieved context to answer the question. If you don't know the answer, just say that you don't know. 
+Question: {question} 
+Context: {context} 
+Answer:"""
 
 # Define a new chain to return both the answer and the sources
 qa_chain_with_sources = (
@@ -52,13 +55,14 @@ qa_chain_with_sources = (
 # Function to call a RAG LLM query
 def rag_query(query, history):
     # Invoke the chain
-    response = qa_chain_with_sources.invoke(query)
+    r = qa_chain_with_sources.invoke(query)
     
-    answer = response["answer"]
-    unique_sources = list(set(response["sources"]))
+    answer = r["answer"]
+    unique_sources = list(set(r["sources"]))
 
     # Print answers + sources
     output = f"Answer: {answer}\n\nSources:\n" + "\n".join(unique_sources)
+
     return output