attention_with_linear_biases

★ 0 updated 4y ago ⑂ fork

Code for our ALiBi method for transformer language models.

No plain-English explanation yet — one is being written right now. Check back in a minute.