"-1" dimension breaks if no bbox in crop or image

[londumas]

In other type of dataset than COCO, having no bbox in a crop is likely. The "-1" dim in a reshape is imposible then.

Change this:

devit/lib/regionprop.py

Line 79 in cfd0e0e

pred_roi_mask = mask_h.reshape(N, -1, 1) * mask_w.reshape(N, 1, -1)

to this:

pred_roi_mask = mask_h.reshape(N, mask_h.shape[-1], 1) * mask_w.reshape(N, 1, mask_w.shape[-1])

Same there:
gt_roi_mask = mask_h.reshape(N, mask_h.shape[-1], 1) * mask_w.reshape(N, 1, mask_w.shape[-1])

and there:
mask_logits = mask_logits.view(bs, np.prod(mask_logits.shape[1:])) / self.temperature # N x K2

The error is what follows:

RuntimeError: cannot reshape tensor of 0 elements into shape [0, -1, 1] because the unspecified dimension size -1 can be any value and is ambiguous

[liniceyo]

Hello, I encountered a similar issue while performing inference on a custom dataset. Do you have any solutions to resolve this? Thank you very much for your response!

File "/home/com/anaconda3/envs/devit/lib/python3.9/site-packages/torch/multiprocessing/spawn.py", line 69, in _wrap
fn(i, *args)
File "/media/A/code/lyl/devit-main/tools/../detectron2/engine/launch.py", line 125, in _distributed_worker
main_func(*args)
File "/media/A/code/lyl/devit-main/tools/train_net.py", line 201, in main
return trainer.train()
File "/media/A/code/lyl/devit-main/tools/../detectron2/engine/defaults.py", line 496, in train
super().train(self.start_iter, self.max_iter)
File "/media/A/code/lyl/devit-main/tools/../detectron2/engine/train_loop.py", line 150, in train
self.after_step()
File "/media/A/code/lyl/devit-main/tools/../detectron2/engine/train_loop.py", line 180, in after_step
h.after_step()
File "/media/A/code/lyl/devit-main/tools/../detectron2/engine/hooks.py", line 386, in after_step
self._do_eval()
File "/media/A/code/lyl/devit-main/tools/../detectron2/engine/hooks.py", line 359, in _do_eval
results = self._func()
File "/media/A/code/lyl/devit-main/tools/../detectron2/engine/defaults.py", line 465, in test_and_save_results
self._last_eval_results = self.test(self.cfg, self.model)
File "/media/A/code/lyl/devit-main/tools/../detectron2/engine/defaults.py", line 618, in test
results_i = inference_on_dataset(model, data_loader, evaluator)
File "/media/A/code/lyl/devit-main/tools/../detectron2/evaluation/evaluator.py", line 159, in inference_on_dataset
outputs = model(inputs)
File "/home/com/anaconda3/envs/devit/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1194, in _call_impl
return forward_call(*input, **kwargs)
File "/home/com/anaconda3/envs/devit/lib/python3.9/site-packages/torch/nn/parallel/distributed.py", line 1040, in forward
output = self._run_ddp_forward(*inputs, **kwargs)
File "/home/com/anaconda3/envs/devit/lib/python3.9/site-packages/torch/nn/parallel/distributed.py", line 1000, in _run_ddp_forward
return module_to_run(*inputs[0], **kwargs[0])
File "/home/com/anaconda3/envs/devit/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1194, in _call_impl
return forward_call(*input, **kwargs)
File "/media/A/code/lyl/devit-main/tools/../detectron2/modeling/meta_arch/devit.py", line 1198, in forward
inter_dist_emb = other_classes.reshape(bs * num_active_classes, -1, self.roialign_size, self.roialign_size)
RuntimeError: cannot reshape tensor of 0 elements into shape [0, -1, 7, 7] because the unspecified dimension size -1 can be any value and is ambiguous