Mariem MOUNIR
Mariem MOUNIR
Im facing the same problem. Do u have any solution ?
Waiting for this to be merged
Did you find a solution for this issue, I'm having the same once I upgraded llama_cpp_python
but relevant nodes based on what? don't we calculate the cos distance and get the top 10 closest element
so if I have 10 as my top k and just 5 were returned then the other 5 were duplicated?