Re: [ot][spam][crazy][data] transformer model 'attention' improvement

Sun, 30 Jan 2022 13:42:48 -0800 

commit 1aa9022c6cd96a93e3c83801ce4fddb5b650ba45 (HEAD ->
memory-efficient-attention, xloem/memory-efficient-attention)
Author: xloem <[email protected]>
Date:   Sun Jan 30 21:41:34 2022 +0000

    draft of chunking without returning attentions. currently fails
due to passing wrong dimensions or data.

Reply via email to