Open ThiagoViek opened 4 years ago
@ThiagoViek Did you solve this problem? I need help with this.
@nataliameira It had to do with lack of RAM on the Jetson. If you are having this problem on a machine with enough memory, try reducing your batch size.
I am currently using a NVIDIA Jetson TX2 with Ubuntu 18.04 to train a Mask R-CNN (Matterport implementation https://github.com/matterport/Mask_RCNN) with my custom dataset. When I try to load the
mask_rcnn_coco.h5
, the following error comes up multiple times:The same error appears in many topics when training the model, but here I am just trying to load the pre-trained weights. I am not understanding why only a small piece of memory is being employed.