News for hackers. Stuff that matters.
Re:Show HN: less than 650 LOC trainable GPT only using NumPy... by PartiallyTyped on Friday November 17, 2023 @01:14PM (#38307565)
The most interesting thing in this whole saga is that decoder only models (aka causal transformers like GPT) are as effective as they are.

There are no comments on this comment yet.