>Brother pdf anon
Stole a Python script online and modified it for my use. I kind of understand how this stuff works now. >Still just barely installed the dependencies like conda, pytorch and docker though
I can now summarize PDFs to the their top 50 relevant chunks which are chosen by relevance according to gemma3:1B
>Any ideas on how to integrate this to sillytavern?
Recommend your favourite small models to summarize PDFs, so far i got;
>Qwen2.5 VL 7B
>Llama 3.2 11B
>Gemma 3:1b as stated