efficientvit icon indicating copy to clipboard operation
efficientvit copied to clipboard

Published 20 hours ago • verified publisher icon mit-han-lab

sam inference using onnx is much slower than trt inference

Open superran1 opened this issue 7 months ago • 0 comments

When infer with trt, sam-l2 can reach a speed of about 15ms, but when infer with onnx, sam-l2 can reach a speed of about 2s

superran1 avatar Jul 25 '24 08:07 superran1