RuntimeError in train eval

[Vio1etovo]

I got an error when evaluating the training set, but this error did not appear in train_half and test. I don't know why, can you answer it?

Error:
File "main_track.py", line 372, in
main(args)
File "main_track.py", line 291, in main
phase='eval', det_val=args.det_val)
File "/home/zbx/Anaconda3/envs/TransTrack-main/lib/python3.6/site-packages/torch/autograd/grad_mode.py", line 15, in decorate_context
return func(*args, **kwargs)
File "/media/zbx/hdd1/Experiments/MOT/TransTrack/engine_track.py", line 133, in evaluate
outputs, pre_embed = model(samples, pre_embed)
File "/home/zbx/Anaconda3/envs/TransTrack-main/lib/python3.6/site-packages/torch/nn/modules/module.py", line 550, in call
result = self.forward(*input, **kwargs)
File "/media/zbx/hdd1/Experiments/MOT/TransTrack/models/deformable_detrtrack_test.py", line 149, in forward
srcs.append(self.combine(torch.cat([self.input_projl, self.input_projl], dim=1)))
RuntimeError: Sizes of tensors must match except in dimension 2. Got 96 and 80