how to change prompt format for zephyr 7b beta model in llamacpp

Hello everyone using mistral 7b instruct as my llm but when applying it into chat engine got incomplete response as compared to llama2 llm model

I am using titleextractor for extracting metadata by passing a document which contains 20 different pdf but i am getting same title for all the PDFs

Hey everyone i am having around 20k pdf files with each of 2-4 pages, i am using MetadataExtractor to extract metadata using llama2 llm but my kernel got crashed how to resolve it??

"What are the key distinctions between using 'as_chat_engine' and 'as_query_engine' in terms of the responses they generate when the same question is asked?"

How can i pass key value in the ExactMatchFilter dynamics means if a user si using my rag chatbot in that case how can i defined my key and value which will be relevant to query

What are the default values of parameters when using llama2 llm such as temperature,context windows

Hello,is there any way in llamaindex to find llama.cpp is running on gpu

you guys did fabulous on query engine , now you need to shift your focus on chatengine because implement it in production is very painful

i have implemented rag using llama-cpp-python with mistral7b openorca model but response time is too high although the api is hosted on sever which has 2 nvidia gpu RTX a4000 . can someone help me out

I am using RAGs with llamacpp as a llm but getting this error

Hey everyone why i am getting empty response when running query engine on a "doc_summary_index"

I have developed a rag system now i want to integrate it on my website as a chatbot. What are the ways in which i can do that for example - i am thinking of creating an API in django but problem is how can i return the response in a streaming mode

Why i am getting these warnings after updating llamaindex version i.e. 0.8.5.post1

Hello folks i am new in rag so can u help me out i have build a rag system using only 2 text file and weaviate vector database. I am getting pretty decent response but it usually take approx 2 min to get the whole response now what should i do to decrease the response time because i need to scale this system for 1000k text files??