Dustin Dannenhauer on Nostr: Maybe AGI will be just one big NN architecture in the end… ...
Maybe AGI will be just one big NN architecture in the end…
https://blog.normalcomputing.ai/posts/2023-09-12-supersizing-transformers/supersizing-transformers.html
I’m still curious what the limitations of the memory cache is here and whether you can keep adding to it without having to retrain it if the cache changes.
https://blog.normalcomputing.ai/posts/2023-09-12-supersizing-transformers/supersizing-transformers.html
I’m still curious what the limitations of the memory cache is here and whether you can keep adding to it without having to retrain it if the cache changes.