Official post on Mixtral 8x7B: https://t.co/ce0ZjHhLVn Official PR into vLLM shows the inference code: https://t.co/vJbmDG9RhG New HuggingFace explainer on MoE very nice: https://t.co/lTaNCONUeI In naive decoding, performance of a bit above 70B (Llama 2), at inference speed… https://t.co/WE9rp74c3D https://t.co/OMSTfYXVsE
— Andrej Karpathy (@karpathy) Dec 11, 2023
from Twitter https://twitter.com/karpathy
December 11, 2023 at 04:38PM
via IFTTT
No comments:
Post a Comment