Kenzo Lobos Tsunekawa

@kminoda The relevant people don't think it is relevant so it being left-as-is for now (we are starting to use 120deg lenses so it may become relevant in the future)

@The-wind-rises-2023 We experience that phenomena not only on Jetson devices but on other ECUs, desktops, etc. Usually you need to tune your dds and network adapter settings. This not really...

Mostly we use this https://autowarefoundation.github.io/autoware-documentation/main/installation/additional-settings-for-developers/ And tune as needed depending on our sensors. Other than that, MTU size on the interface.

I think you meant to mention @The-wind-rises-2023

@amadeuszsz I added transfusion to the other universe launchers. Just for the duration of the review, I set transfusion as the default model. However, before merging we should return it...

@amadeuszsz While doing other task that was asked of me, I was looking at the whole perception task and saw the preprocessing being done in cpu in centerpoint. I kind...

@amadeuszsz After reading your kernels seriously this time, I realized that you went to great lengths to make the operations general. Usually, when writing kernels (cpu or gpu), you want...

A while ago (how time flies), it was decided that we would migrate towards this way of using parameters: https://github.com/orgs/autowarefoundation/discussions/3371 While not all the nodes have implemented this (in particular...

@amadeuszsz Reading the preprocessing -> inference -> postprocessing part, I see several syncs that are needed to check the error code and abort midway when needed. Are all of these...

@Sophie2216 Hi, I have observed that behavior a few times last year, but I have not been able to reproduce it since. I would recommend you to disable the drawings...