search

opendilab / LMDrive

[CVPR 2024] LMDrive: Closed-Loop End-to-End Driving with Large Language Models
Apache License 2.0
526 stars 48 forks source link

inference frequency #32

Open PzWHU opened 3 months ago

PzWHU commented 3 months ago

Hi! Thanks for your excellent work. What is the frequency of model inference output on A100？ Thanks!

PzWHU commented 3 months ago

or on 3090

deepcs233 commented 3 months ago

Hi!

With 3090, a frame in the simulator usually takes about 0.15 seconds.