Problem with input with spatial dimensions of [256] and output size of (138, 138)

[yu123569]

I'm run the train.py by custom dataset successfully on colab/colab pro. in March 6 2023.
I start my script to install everything I need, and I got this error yeserday :

"Traceback (most recent call last):
File "train.py", line 707, in
train(0, args=args)
File "train.py", line 433, in train
losses = backward_and_log(run_name, net_outs, targets, masks, num_crowds)
File "train.py", line 325, in backward_and_log
losses = criterion(out, targets, masks, num_crowds)
File "/usr/local/lib/python3.8/dist-packages/torch/nn/modules/module.py", line 1194, in _call_impl
return forward_call(*input, **kwargs)
File "/content/gdrive/MyDrive/ColabNotebooks/yolact_edge-master/yolact_edge/layers/modules/multibox_loss.py", line 159, in forward
losses.update(self.lincomb_mask_loss(pos, idx_t, loc_data, mask_data, priors, proto_data, masks, gt_box_t, inst_data))
File "/content/gdrive/MyDrive/ColabNotebooks/yolact_edge-master/yolact_edge/layers/modules/multibox_loss.py", line 442, in lincomb_mask_loss
downsampled_masks = F.interpolate(masks[idx].unsqueeze(0), (mask_h, mask_w),
File "/usr/local/lib/python3.8/dist-packages/torch/nn/functional.py", line 3866, in interpolate
raise ValueError(
ValueError: Input and output must have the same number of spatial dimensions, but got input with spatial dimensions of [256] and output size of (138, 138). Please provide input tensor in (N, C, d1, d2, ...,dK) format and output size in (o1, o2, ...,oK) format."

I had try to downgrade:
CUDA 11.8 -> 11.6 (follow with correct version of cudnn and TensorRT)
Python 3.9-> 3.8
The tensorflow-gpu(2.9.2) and pytorch(1.13.1+cu116) version just as before.
Total 4 combination of version, but still got the same error.

I guess it is related with some package verison.
Can I get any one pip list if your train.py still working?

[yu123569]

Find out the error come from train.py line 762 : 'use_gt_bboxes': False ->True
How does it happen?