AttributeError: 'tuple' object has no attribute 'shape'

[sonnystalin]

Hi, I'm having trouble using the code to train on my own dataset. The problem occurs with the Gaussian_Diffusion module.

The question is as follows：

Traceback (most recent call last): File "scripts/segmentation_train.py", line 118, in main() File "scripts/segmentation_train.py", line 70, in main TrainLoop( File "E:\MedSegDiff-Spine-main.\guided_diffusion\train_util.py", line 180, in run_loop self.run_step(batch, cond) File "E:\MedSegDiff-Spine-main.\guided_diffusion\train_util.py", line 201, in run_step sample = self.forward_backward(batch, cond) File "E:\MedSegDiff-Spine-main.\guided_diffusion\train_util.py", line 231, in forward_backward losses1 = compute_losses() File "E:\MedSegDiff-Spine-main.\guided_diffusion\gaussian_diffusion.py", line 650, in training_losses_segmentation terms["loss"] = self._vb_terms_bpd( File "E:\MedSegDiff-Spine-main.\guided_diffusion\gaussian_diffusion.py", line 614, in _vb_terms_bpd out = self.p_mean_variance( File "E:\MedSegDiff-Spine-main.\guided_diffusion\respace.py", line 90, in p_mean_variance return super().p_mean_variance(self._wrap_model(model), *args, **kwargs) File "E:\MedSegDiff-Spine-main.\guided_diffusion\gaussian_diffusion.py", line 267, in p_mean_variance model_mean = self.q_posterior_mean_variance( File "E:\MedSegDiff-Spine-main.\guided_diffusion\gaussian_diffusion.py", line 212, in q_posterior_mean_variance assert x_start.shape == x_t.shape AttributeError: 'tuple' object has no attribute 'shape'

The format of x_start is： x_start type: <class 'tuple'>, value: (tensor([[[[0., 0., 0., ..., 0., 0., 0.], [0., 0., 0., ..., 0., 0., 0.], [0., 0., 0., ..., 0., 0., 0.], ..., [0., 0., 0., ..., 0., 0., 0.], [0., 0., 0., ..., 0., 0., 0.], [0., 0., 0., ..., 0., 0., 0.]],

     [[0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      ...,
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.]]],


    [[[0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      ...,
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.]],

     [[0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      ...,
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.]]],


    [[[0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      ...,
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.]],

     [[0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      ...,
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.]]],


    ...,


    [[[0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      ...,
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.]],

     [[0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      ...,
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.]]],


    [[[0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      ...,
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.]],

     [[0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      ...,
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.]]],


    [[[0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      ...,
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.]],

     [[0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      ...,
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.],
      [0., 0., 0.,  ..., 0., 0., 0.]]]], device='cuda:0',
   grad_fn=<ConvolutionBackward0>), tensor([[[[0.5846, 0.4764, 0.4932,  ..., 0.5746, 0.5637, 0.5700],
      [0.5851, 0.6393, 0.7173,  ..., 0.5075, 0.5434, 0.4702],
      [0.5372, 0.5826, 0.8107,  ..., 0.5443, 0.6231, 0.4982],
      ...,
      [0.6365, 0.5709, 0.5678,  ..., 0.4498, 0.5057, 0.4866],
      [0.5337, 0.5445, 0.4393,  ..., 0.3687, 0.5252, 0.5495],
      [0.6406, 0.6942, 0.6227,  ..., 0.6035, 0.6418, 0.6114]]],


    [[[0.5541, 0.5860, 0.6421,  ..., 0.6785, 0.6934, 0.6377],
      [0.5828, 0.5905, 0.5781,  ..., 0.6483, 0.6696, 0.6074],
      [0.6532, 0.5950, 0.5979,  ..., 0.6306, 0.6737, 0.6274],
      ...,
      [0.6714, 0.6825, 0.6744,  ..., 0.5590, 0.5939, 0.5054],
      [0.6674, 0.6215, 0.6316,  ..., 0.4703, 0.4789, 0.4470],
      [0.6571, 0.6312, 0.6461,  ..., 0.5525, 0.5777, 0.5481]]],


    [[[0.5368, 0.5429, 0.5488,  ..., 0.4852, 0.4854, 0.4710],
      [0.5555, 0.5635, 0.6236,  ..., 0.5658, 0.5751, 0.4788],
      [0.5509, 0.5510, 0.6257,  ..., 0.5238, 0.5754, 0.5233],
      ...,
      [0.7045, 0.6531, 0.5366,  ..., 0.5109, 0.5114, 0.4738],
      [0.6715, 0.6297, 0.3976,  ..., 0.4826, 0.4157, 0.4372],
      [0.6479, 0.6427, 0.5018,  ..., 0.5952, 0.4995, 0.5293]]],


    ...,


    [[[0.5558, 0.5454, 0.6055,  ..., 0.5731, 0.5199, 0.5411],
      [0.5262, 0.6009, 0.7195,  ..., 0.5634, 0.5865, 0.5473],
      [0.5882, 0.6384, 0.6448,  ..., 0.4815, 0.6975, 0.4595],
      ...,
      [0.4299, 0.8450, 0.5439,  ..., 0.5943, 0.4213, 0.3321],
      [0.4318, 0.6034, 0.6760,  ..., 0.6217, 0.4202, 0.5185],
      [0.7269, 0.6352, 0.6566,  ..., 0.6751, 0.4266, 0.6503]]],


    [[[0.3344, 0.4871, 0.4382,  ..., 0.6347, 0.6236, 0.5605],
      [0.4082, 0.6057, 0.2145,  ..., 0.5594, 0.7679, 0.4599],
      [0.5393, 0.1673, 0.9391,  ..., 0.4425, 0.6835, 0.6310],
      ...,
      [0.7584, 0.6826, 0.5471,  ..., 0.8120, 0.3843, 0.5641],
      [0.5432, 0.5330, 0.3213,  ..., 0.4990, 0.4184, 0.5196],
      [0.6815, 0.6538, 0.7469,  ..., 0.7142, 0.5969, 0.6261]]],


    [[[0.6194, 0.7726, 0.6500,  ..., 0.6507, 0.5600, 0.5233],
      [0.7337, 0.6887, 0.3808,  ..., 0.7279, 0.5641, 0.5267],
      [0.5793, 0.3779, 0.3818,  ..., 0.9037, 0.8900, 0.2796],
      ...,
      [0.7714, 0.9128, 0.7506,  ..., 0.6294, 0.4318, 0.4752],
      [0.7695, 0.6399, 0.5966,  ..., 0.6750, 0.3901, 0.4319],
      [0.6037, 0.6148, 0.6483,  ..., 0.5312, 0.5273, 0.5470]]]],
   device='cuda:0', grad_fn=<SigmoidBackward0>))

Would you please help me solve this problem？Thank you! @WuJunde