Hey, love the plugin. One thing I'd like to see is to have the option for the summarizer to work in batches whenever the current context's token length hits a certain set threshold set by the user. I'm not sure if it's 100% feasible to implement, but I think it would serve as a pretty effective way to batch summarize old context when it actually becomes a problem (i.e. when it's approaching your model's max allowed context limit) instead of the current trigger being an arbitrary number of turns. Would be nice if it could be looked into.
Hey, love the plugin. One thing I'd like to see is to have the option for the summarizer to work in batches whenever the current context's token length hits a certain set threshold set by the user. I'm not sure if it's 100% feasible to implement, but I think it would serve as a pretty effective way to batch summarize old context when it actually becomes a problem (i.e. when it's approaching your model's max allowed context limit) instead of the current trigger being an arbitrary number of turns. Would be nice if it could be looked into.