stack expects each tensor to be equal size, but got [531, 674, 3] at entry 0 and [380, 885, 3] at entry 1

[dididichufale]

I have no way to solve this kind of problem. This kind of problem is usually caused by data processing, but I can't find the error. What can I do?
D:\annaconda3\envs\aimbot_env\python.exe J:/stereo-transformer-main/stereo-transformer/main.py
number of params in backbone: 1,050,800
number of params in transformer: 797,440
number of params in tokenizer: 503,728
number of params in regression: 161,843
0%| | 0/11195 [00:00<?, ?it/s]Start training
Epoch: 0
0%| | 0/11195 [00:03<?, ?it/s]
Traceback (most recent call last):
File "J:/stereo-transformer-main/stereo-transformer/main.py", line 263, in
main(args_)
File "J:/stereo-transformer-main/stereo-transformer/main.py", line 234, in main
args.clip_max_norm, amp)
File "J:\stereo-transformer-main\stereo-transformer\utilities\train.py", line 30, in train_one_epoch
for idx, data in enumerate(tbar):
File "D:\annaconda3\envs\aimbot_env\lib\site-packages\tqdm\std.py", line 1180, in iter
for obj in iterable:
File "D:\annaconda3\envs\aimbot_env\lib\site-packages\torch\utils\data\dataloader.py", line 521, in next
data = self._next_data()
File "D:\annaconda3\envs\aimbot_env\lib\site-packages\torch\utils\data\dataloader.py", line 1203, in _next_data
return self._process_data(data)
File "D:\annaconda3\envs\aimbot_env\lib\site-packages\torch\utils\data\dataloader.py", line 1229, in _process_data
data.reraise()
File "D:\annaconda3\envs\aimbot_env\lib\site-packages\torch_utils.py", line 434, in reraise
raise exception
RuntimeError: Caught RuntimeError in DataLoader worker process 0.
Original Traceback (most recent call last):
File "D:\annaconda3\envs\aimbot_env\lib\site-packages\torch\utils\data_utils\worker.py", line 287, in _worker_loop
data = fetcher.fetch(index)
File "D:\annaconda3\envs\aimbot_env\lib\site-packages\torch\utils\data_utils\fetch.py", line 52, in fetch
return self.collate_fn(data)
File "D:\annaconda3\envs\aimbot_env\lib\site-packages\torch\utils\data_utils\collate.py", line 74, in default_collate
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "D:\annaconda3\envs\aimbot_env\lib\site-packages\torch\utils\data_utils\collate.py", line 74, in
return {key: default_collate([d[key] for d in batch]) for key in elem}
File "D:\annaconda3\envs\aimbot_env\lib\site-packages\torch\utils\data_utils\collate.py", line 64, in default_collate
return default_collate([torch.as_tensor(b) for b in batch])
File "D:\annaconda3\envs\aimbot_env\lib\site-packages\torch\utils\data_utils\collate.py", line 56, in default_collate
return torch.stack(batch, 0, out=out)
RuntimeError: stack expects each tensor to be equal size, but got [531, 674, 3] at entry 0 and [380, 885, 3] at entry 1

[Guzaiwang]

Yes, I have met the same problem.
This is because of the random crop augmentation in the data loader process.
If you want to set the batch size > 2, you can do the random crop process in the function ( forward_pass ) rather than in the definition of getitem.

[Guzaiwang]

When you set the batch size > 2, another problem may occur.
It is because the batch size equals to 1 in the original code.
You can see line 109 in the transform.py file.

Finally, the code could run and work well.