BUG: Evaluate model using COCO dataset

Open adrianosantospb opened this issue 5 years ago • 0 comments

Thank you for share your repository and implementation. I'm training a model using my own dataset. The training process seems to work well, but when the model calls the evaluation function, I get this error:

File "/home/workstation/anaconda3/envs/efficientdet/lib/python3.7/site-packages/torch/nn/parallel/scatter_gather.py", line 55, in gather_map return Gather.apply(target_device, dim, *outputs) File "/home/workstation/anaconda3/envs/efficientdet/lib/python3.7/site-packages/torch/nn/parallel/_functions.py", line 54, in forward assert all(map(lambda i: i.is_cuda, inputs))

This error occurs in this call:

scores, labels, boxes = model(data['img'].permute(2, 0, 1).cuda().float().unsqueeze(dim=0)) Some idea?

Jul 10 '20 16:07 adrianosantospb