Elon Musk Archive
Elon Musk
Elon Musk
@elonmusk
En respuesta a @O42n2
Tesla does INT8 inference. Way more efficient than FP16, but took us a lot of effort to overcome quantization errors.
19:54 · 28 de febrero de 2023 · 137,1 mil visualizaciones
71
80
1,3 mil
Open on
@O42nl @kirillgroshkov Tesla does INT8 inference. Way more efficient than FP16, but took us a lot of effort to overcome quantization errors. — Elon Musk Archive