Pretrained Model for Binary_Event_Image

82magnolia / n_imagenet

Official PyTorch implementation of N-ImageNet: Towards Robust, Fine-Grained Object Recognition with Event Cameras (ICCV 2021)

GNU General Public License v3.0

52 stars 5 forks source link

Pretrained Model for Binary_Event_Image #8

Open ldkong1205 opened 7 months ago

ldkong1205 commented 7 months ago

Hi @82magnolia, thanks for your great contribution!

I encountered the following error when testing the Binary_Event_Image model downloaded from your Google Drive at this link:

RuntimeError: Given groups=1, weight of size [64, 2, 14, 14], expected input[256, 1, 224, 224] to have 2 channels, but got 1 channels instead

Could you please take a look? Thanks.

82magnolia commented 6 months ago

Hi! Sorry for the late reply. I suspect that the problem is with the input event representation: if I remember correctly binary_event_image model is trained on a two channel event image (with each channel representing positive and negative events respectively). Can you share the exact command you used when you ran into the error?