Hello everyone, I thought all public web-based KV cache calculators kinda suck.. so I decided to create one I would like to use myself - KVANTA https://kvanta.vcerny.cz It should support any LLM/VLM from Hugging Face, if not let me know! (also, it's Apache 2.0) https://preview.redd.it/rk8i48ftva3h1.
