Comment on Ring attention with blockwise transformers for near-infinite contextparentComments−ivalm2yWhy do you think so? Many people tried neuroinspired stuff and had no results. Entire field of comp neurosci produced what exactly over the past decade? People invest in transformers because they work while everything else so far doesn’t.
Comments
Why do you think so? Many people tried neuroinspired stuff and had no results. Entire field of comp neurosci produced what exactly over the past decade? People invest in transformers because they work while everything else so far doesn’t.