assertion error assert x_t.shape == eps.shape

[worthlessFella]

Hi I have trained the segmentation code on dataset ISIC. But when I am running sample I am getting the following error. I have used all the default parameters suggested and have not made any changes to the code except few print statement for debugging:

[W socket.cpp:663] [c10d] The client socket has failed to connect to [::ffff:127.0.1.1]:56131 (system error: 10049 - The requested address is not valid in its context.).
Logging to ......\datasets\MedSegDiff\sample
creating model and diffusion...
sampling...
sample_fn False <bound method GaussianDiffusion.p_sample_loop_known of <guided_diffusion.respace.SpacedDiffusion object at 0x000002C1AD337E50>>
p_sample_loop_known
no dpm-solver
p_sample torch.Size([1, 4, 64, 64])
x torch.Size([1, 4, 64, 64])
in _predict_xstart_from_eps: x.shape, eps.shape torch.Size([1, 1, 64, 64]) torch.Size([1, 2, 64, 64])
Traceback (most recent call last):
File "C:\Users\bdebnath\data\myprojects\MedSegDiff\scripts\segmentation_sample.py", line 207, in
main()
File "C:\Users\bdebnath\data\myprojects\MedSegDiff\scripts\segmentation_sample.py", line 117, in main
sample, x_noisy, org, cal, cal_out = sample_fn(
File "C:\Users\bdebnath\data\myprojects\MedSegDiff\guided_diffusion\gaussian_diffusion.py", line 569, in p_sample_loop_known
for sample in self.p_sample_loop_progressive(
File "C:\Users\bdebnath\data\myprojects\MedSegDiff\guided_diffusion\gaussian_diffusion.py", line 654, in p_sample_loop_progressive
out = self.p_sample(
File "C:\Users\bdebnath\data\myprojects\MedSegDiff\guided_diffusion\gaussian_diffusion.py", line 447, in p_sample
out = self.p_mean_variance(
File "C:\Users\bdebnath\data\myprojects\MedSegDiff\guided_diffusion\respace.py", line 90, in p_mean_variance
return super().p_mean_variance(self._wrap_model(model), *args, **kwargs)
File "C:\Users\bdebnath\data\myprojects\MedSegDiff\guided_diffusion\gaussian_diffusion.py", line 329, in p_mean_variance
self._predict_xstart_from_eps(x_t=x, t=t, eps=model_output)
File "C:\Users\bdebnath\data\myprojects\MedSegDiff\guided_diffusion\gaussian_diffusion.py", line 350, in _predict_xstart_from_eps
assert x_t.shape == eps.shape
AssertionError

[FionaYeager]

I also encountered the same problem. Have you resolved this bug now？

[luisfernandes9]

I am having the same problem as well

[linyuanwang]

I have encountered the same problem. Has anyone solved it ?