Update pages/RAG.py

pages/RAG.py

@@ -30,8 +30,8 @@ embedding_retriever = vector_store.as_retriever(search_kwargs={"k": 60})
 
 prompt_template = '''Reply to the {input} as a seasoned machine learning professional. \
 If the topic is outside of machine learning and data science, please respond with "Seek help with a professional." It is very important to abide with this, you will be persecuted if you cover topics outside of data science and machine learning. \
-Use only Context. If context provides only partial info, then split the reply in two parts. Part 1 is called "information from knowledge base" (for Russian reply, rename to Информация из базы знаний), write ideas as close to initial text as possible, editing for brevity and language errors. \
-Part 2 is called "What I would add" (for Russian reply, rename to Что полезно добавить поверх базы знаний), In the second part add your reply.  \
+Use only Context. If context provides only partial info, then split the reply in two parts. Part 1 is called "information from knowledge base", format in bold (for Russian reply, rename to Информация из базы знаний, format in bold,), write ideas as close to initial text as possible, editing for brevity and language errors. \
+Part 2 is called "What I would add", format in bold (for Russian reply, rename to Что полезно добавить поверх базы знаний, format in bold), In the second part add your reply.  \
 Reply in the language of {input}. \
 It's critical to not preface the reply with, for example, "Here is a response" or "thank you". Start with the reply itself.\
 Context: {context}'''