Elon Musk Archive
Elon Musk
Elon Musk
@elonmusk
En respuesta a @samiralyateem
There are a lot of pointless zeroes in FP32 neural nets. You can chop off 16 of the 32 bits without losing meaningful precision. This works well with neural nets, but not with regular computing, which expects extreme precision.
23:34 · 20 de septiembre de 2020 · 0 visualizaciones
27
17
491
Open on
@samiralyateem @ID_AA_Carmack There are a lot of pointless zeroes in FP32 neural nets. You can chop off 16 of the 32 bits without losing meaningful precision. This works well with neural nets, but not with regular computing, which expects extreme precision. — Elon Musk Archive