search

NVIDIA-AI-IOT / jetracer

An autonomous AI racecar using NVIDIA Jetson Nano
MIT License
1.06k stars 319 forks source link

Out of memory when training with other models #68

Closed RogerZ1128 closed 2 years ago

RogerZ1128 commented 4 years ago

Hi， I've successfully run the training and roadfollowing with default resnet18 and I noticed that there are some more models to choose. For some need, I'd like to try other models that take less system memory. As far as I know, Alexnet and squeezenet are lighter than resnet18. However, when I run training with these models, system will be out of memory and crashed. What's the problem？Did I miss some settings? Thanks.

diamondbarcode commented 4 years ago

hi RogerZ1128 is there any tip on how to get this road following work? i try mine and i could not get it to work properly ..

RogerZ1128 commented 4 years ago

hi RogerZ1128 is there any tip on how to get this road following work? i try mine and i could not get it to work properly ..

What's the matter?

diamondbarcode commented 4 years ago

I mean how you get it to follow road properly with resnet18? I could not get it to do work properly.

tankhunter8192 commented 4 years ago

try to higher the amount of swap memory. Easiest way -> jtop/jetsonstats