Banaxi PRO

individually please, id also like you guys using NSA (or bga lol), also as a tip you need to do more overtraining

That's true

i already deleted the blog

well should be because with local it can keep nearby syntax and code structure etc, and cold is like compressed summary so not fully gone

well from the some i know, bga routes blocks while dsa routes tokens, and doesent have the cold and local memory

i wanted try to make a model architecture and dsa wasnt that good i never knew about nsa until now and yes i used ai to make the arch yes only thought better than dsa because i NEVER Knew ab out nsa idk what you want me to do you keep now i already know and yeah ok ill stop making new architectures and just use preexissting ones i did read tge bkog i just didnt know about nsa only compared it to dsa

@DedeProGames dont try

we delaying bananamind 3 😭

ok so thats what you want to do? ok

so what do you want me to do? take it down?

i didnt know about NSA only dsa

then what the hell is this:

the cold reccurent state of BGA is not same as NSA's compression branh, NSA's compressed tokens are another attention branch bga cold state is a recurrent summary carried forward

just stop lying.....

i looked it up then bhen you mentioned NSA

i didnt know about NSA only dsa

thanks for letting me know that, dude. I thought this guy was doing actual research all of a sudden.

yea yea wait for BM3X

the cold reccurent state of BGA is not same as NSA's compression branh, NSA's compressed tokens are another attention branch bga cold state is a recurrent summary carried forward

Well this provides better quality because of the nearby memory and the others memory forms.