NotImplementedError: Input Error: Only 3D, 4D and 5D input Tensors supported (got 3D) for the modes: nearest | linear | bilinear | bicubic | trilinear (got bicubic)

[hfarhidzadeh]

Hi @shunsukesaito

I could run the code on colab on my test data and worked fine. I am trying to run it on gpu machine getting this following error.

Traceback (most recent call last):
File "/home/hamid_farhid/anaconda3/envs/PIFuHD/lib/python3.7/runpy.py", line 193, in _run_module_as_main
"main", mod_spec)
File "/home/hamid_farhid/anaconda3/envs/PIFuHD/lib/python3.7/runpy.py", line 85, in _run_code
exec(code, run_globals)
File "/home/hamid_farhid/PIFuHD/pifuhd/apps/simple_test.py", line 30, in
reconWrapper(cmd, args.use_rect)
File "/home/hamid_farhid/PIFuHD/pifuhd/apps/recon.py", line 220, in reconWrapper
recon(opt, use_rect)
File "/home/hamid_farhid/PIFuHD/pifuhd/apps/recon.py", line 210, in recon
gen_mesh(opt.resolution, netMR, cuda, test_data, save_path, components=opt.use_compose)
File "/home/hamid_farhid/PIFuHD/pifuhd/apps/recon.py", line 38, in gen_mesh
net.filter_global(image_tensor_global)
File "/home/hamid_farhid/PIFuHD/pifuhd/lib/model/HGPIFuMRNet.py", line 83, in filter_global
self.netG.filter(images)
File "/home/hamid_farhid/PIFuHD/pifuhd/lib/model/HGPIFuNetwNML.py", line 134, in filter
self.im_feat_list, self.normx = self.image_filter(images)
File "/home/hamid_farhid/anaconda3/envs/PIFuHD/lib/python3.7/site-packages/torch/nn/modules/module.py", line 489, in call
result = self.forward(*input, **kwargs)
File "/home/hamid_farhid/PIFuHD/pifuhd/lib/model/HGFilters.py", line 195, in forward
hg = self._modules'm' + str(i)
File "/home/hamid_farhid/anaconda3/envs/PIFuHD/lib/python3.7/site-packages/torch/nn/modules/module.py", line 489, in call
result = self.forward(*input, **kwargs)
File "/home/hamid_farhid/PIFuHD/pifuhd/lib/model/HGFilters.py", line 117, in forward
return self._forward(self.depth, x)
File "/home/hamid_farhid/PIFuHD/pifuhd/lib/model/HGFilters.py", line 103, in _forward
low2 = self._forward(level - 1, low1)
File "/home/hamid_farhid/PIFuHD/pifuhd/lib/model/HGFilters.py", line 111, in _forward
up2 = F.interpolate(low3, scale_factor=2, mode='bicubic', align_corners=True)
File "/home/hamid_farhid/anaconda3/envs/PIFuHD/lib/python3.7/site-packages/torch/nn/functional.py", line 2459, in interpolate
" (got {})".format(input.dim(), mode))
NotImplementedError: Input Error: Only 3D, 4D and 5D input Tensors supported (got 4D) for the modes: nearest | linear | bilinear | trilinear (got bicubic)
freeglut (foo): failed to open display ''

By referencing this: Error, we should do some changes on input tensor. Any thought on this one?

[shunsukesaito]

Can you make sure you use the latest version of pytorch?

[hfarhidzadeh]

@shunsukesaito Thank you. I updated the version of pytorch. I ran to new issue when I rum demo.sh on remote desktop
Warning: opt is overwritten. test data size: 0 initialize network with normal initialize network with normal generate mesh (test) ... 0it [00:00, ?it/s]

[shunsukesaito]

Seems like files under the folder is not recognized. Do you put PNG/JPG files together with "{file_name}_keypoints.json" or "{file_name}_rect.txt"? Please check if the code works first with the provided sample image.

[shunsukesaito]

I will close the issue. Please reopen if the problem persists.