I ran into cuda OOM error when running Gemma 7b on ollama. My GPU has 8G memory and can run llama2 7b and Mistral 7b without issues. I could run Gemma 7b in ollama CLI though, just not via llamaindex in a rag app.
Hi, does llamaindex support continous indexing with newer data? The getting started code only shows its batch indexing capability (i.e. index all data files in the \data folder)
