Skill
jetson-speculative-decoding
Created by nvidia
Version 392ab8ab2587345a9d86240967f666bd59ac82ca
Overview
Add EAGLE-3 or draft-model speculative decoding to a Jetson vLLM server when TPOT is the bottleneck.
Skill
Created by nvidia
Version 392ab8ab2587345a9d86240967f666bd59ac82ca
Add EAGLE-3 or draft-model speculative decoding to a Jetson vLLM server when TPOT is the bottleneck.