We managed to increase the ctx for whatever llm model that is GGUFED, here the experimental tests: https://www.reddit.com/r/Hugston/