Very cool idea. Can't wait for converted models on HF. | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		olq_plo 9 months ago \| parent \| context \| favorite \| on: TransMLA: Multi-head latent attention is all you n... Very cool idea. Can't wait for converted models on HF.

MichaelMoser123 9 months ago [–]

deepseek-v2,v3,r1 are all using multi-headed attention.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact