TypeError: 'tuple' object is not callable

[cdcseacave]

I get the following error when running bash run.sh:

Setting up PyTorch plugin "gridsample_grad2"... /home/dan/miniconda3/envs/petneus/lib/python3.8/site-packages/torch/functional.py:504: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at ../aten/src/ATen/native/TensorShape.cpp:3483.)
  return _VF.meshgrid(tensors, **kwargs)  # type: ignore[attr-defined]
/home/dan/miniconda3/envs/petneus/lib/python3.8/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
  warnings.warn(
/home/dan/miniconda3/envs/petneus/lib/python3.8/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
  warnings.warn(msg)

  0%|          | 0/100000 [00:00<?, ?it/s]
  0%|          | 0/100000 [00:00<?, ?it/s]
Done.
Hello Wooden
Load data: Begin
Load data: End
Setting up [LPIPS] perceptual loss: trunk [vgg], v[0.1], spatial [off]
Loading model from: /home/dan/miniconda3/envs/petneus/lib/python3.8/site-packages/lpips/weights/v0.1/vgg.pth
Traceback (most recent call last):
  File "exp_runner_pet.py", line 497, in <module>
    runner.train()
  File "exp_runner_pet.py", line 167, in train
    render_out = self.renderer.render(rays_o, rays_d, near, far,
  File "/mnt/d/apps/PET-NeuS/models/renderer_pet.py", line 364, in render
    gradient = self.sdf_network.gradient(pts.reshape(-1, 3)).squeeze().detach()
  File "/mnt/d/apps/PET-NeuS/models/triplane_pet.py", line 159, in gradient
    gradients = torch.autograd.grad(
  File "/home/dan/miniconda3/envs/petneus/lib/python3.8/site-packages/torch/autograd/__init__.py", line 303, in grad
    return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
  File "/home/dan/miniconda3/envs/petneus/lib/python3.8/site-packages/torch/autograd/function.py", line 274, in apply
    return user_fn(self, *args)
  File "/mnt/d/apps/PET-NeuS/third_party/ops/grid_sample.py", line 59, in backward
    grad_input, grad_grid = _GridSample2dBackward.apply(grad_output, input, grid, ctx.padding_mode, ctx.align_corners)
  File "/home/dan/miniconda3/envs/petneus/lib/python3.8/site-packages/torch/autograd/function.py", line 506, in apply
    return super().apply(*args, **kwargs)  # type: ignore[misc]
  File "/mnt/d/apps/PET-NeuS/third_party/ops/grid_sample.py", line 68, in forward
    grad_input, grad_grid = op(grad_output, input, grid, 0, padding_mode, align_corners, output_mask)
TypeError: 'tuple' object is not callable

[emiald]

I meet the same problems and resolve it .You should debug the grid_sample.py and change it. op=torch._C._jit_get_operation('aten::grid_sampler_2d_backword') is a tuple that cause the problems.
Change it to op=torch._C._jit_get_operation('aten::grid_sampler_2d_backword')[0]

[cdcseacave]

thx, that helps a bit, but now I have a new error:

Setting up PyTorch plugin "gridsample_grad2"... /home/dan/miniconda3/envs/petneus/lib/python3.8/site-packages/torch/functional.py:504: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at ../aten/src/ATen/native/TensorShape.cpp:3483.)
  return _VF.meshgrid(tensors, **kwargs)  # type: ignore[attr-defined]
/home/dan/miniconda3/envs/petneus/lib/python3.8/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
  warnings.warn(
/home/dan/miniconda3/envs/petneus/lib/python3.8/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
  warnings.warn(msg)

  0%|          | 0/100000 [00:00<?, ?it/s]
  0%|          | 0/100000 [00:00<?, ?it/s]
Done.
Hello Wooden
Load data: Begin
Load data: End
Setting up [LPIPS] perceptual loss: trunk [vgg], v[0.1], spatial [off]
Loading model from: /home/dan/miniconda3/envs/petneus/lib/python3.8/site-packages/lpips/weights/v0.1/vgg.pth
Traceback (most recent call last):
  File "exp_runner_pet.py", line 497, in <module>
    runner.train()
  File "exp_runner_pet.py", line 167, in train
    render_out = self.renderer.render(rays_o, rays_d, near, far,
  File "/mnt/d/apps/PET-NeuS/models/renderer_pet.py", line 393, in render
    ret_fine = self.render_core(rays_o,
  File "/mnt/d/apps/PET-NeuS/models/renderer_pet.py", line 269, in render_core
    gradients = sdf_network.gradient(pts).squeeze()
  File "/mnt/d/apps/PET-NeuS/models/triplane_pet.py", line 157, in gradient
    y = self.sdf(x)
  File "/mnt/d/apps/PET-NeuS/models/triplane_pet.py", line 153, in sdf
    return self.forward(coordinates)[:, :1]
  File "/mnt/d/apps/PET-NeuS/models/triplane_pet.py", line 97, in forward
    return self.run_model(planes, self.decoder, coordinates.unsqueeze(0), directions, self.rendering_kwargs)
  File "/mnt/d/apps/PET-NeuS/models/triplane_pet.py", line 116, in run_model
    attn_windows = self.attn4(x_windows)
  File "/home/dan/miniconda3/envs/petneus/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
    return forward_call(*args, **kwargs)
  File "/mnt/d/apps/PET-NeuS/models/swin_transformer.py", line 136, in forward
    attn = (q @ k.transpose(-2, -1))
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.50 GiB (GPU 0; 15.99 GiB total capacity; 13.59 GiB already allocated; 0 bytes free; 14.21 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

[YunnaChen]

uld debug th
I also meet this problem, do you solve it?

[520jz]

uld debug th
I also meet this problem, do you solve it?

Maybe, you can change the batch_size

[520jz]

thx, that helps a bit, but now I have a new error:

Setting up PyTorch plugin "gridsample_grad2"... /home/dan/miniconda3/envs/petneus/lib/python3.8/site-packages/torch/functional.py:504: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at ../aten/src/ATen/native/TensorShape.cpp:3483.)
  return _VF.meshgrid(tensors, **kwargs)  # type: ignore[attr-defined]
/home/dan/miniconda3/envs/petneus/lib/python3.8/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.
  warnings.warn(
/home/dan/miniconda3/envs/petneus/lib/python3.8/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG16_Weights.IMAGENET1K_V1`. You can also use `weights=VGG16_Weights.DEFAULT` to get the most up-to-date weights.
  warnings.warn(msg)

  0%|          | 0/100000 [00:00<?, ?it/s]
  0%|          | 0/100000 [00:00<?, ?it/s]
Done.
Hello Wooden
Load data: Begin
Load data: End
Setting up [LPIPS] perceptual loss: trunk [vgg], v[0.1], spatial [off]
Loading model from: /home/dan/miniconda3/envs/petneus/lib/python3.8/site-packages/lpips/weights/v0.1/vgg.pth
Traceback (most recent call last):
  File "exp_runner_pet.py", line 497, in <module>
    runner.train()
  File "exp_runner_pet.py", line 167, in train
    render_out = self.renderer.render(rays_o, rays_d, near, far,
  File "/mnt/d/apps/PET-NeuS/models/renderer_pet.py", line 393, in render
    ret_fine = self.render_core(rays_o,
  File "/mnt/d/apps/PET-NeuS/models/renderer_pet.py", line 269, in render_core
    gradients = sdf_network.gradient(pts).squeeze()
  File "/mnt/d/apps/PET-NeuS/models/triplane_pet.py", line 157, in gradient
    y = self.sdf(x)
  File "/mnt/d/apps/PET-NeuS/models/triplane_pet.py", line 153, in sdf
    return self.forward(coordinates)[:, :1]
  File "/mnt/d/apps/PET-NeuS/models/triplane_pet.py", line 97, in forward
    return self.run_model(planes, self.decoder, coordinates.unsqueeze(0), directions, self.rendering_kwargs)
  File "/mnt/d/apps/PET-NeuS/models/triplane_pet.py", line 116, in run_model
    attn_windows = self.attn4(x_windows)
  File "/home/dan/miniconda3/envs/petneus/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
    return forward_call(*args, **kwargs)
  File "/mnt/d/apps/PET-NeuS/models/swin_transformer.py", line 136, in forward
    attn = (q @ k.transpose(-2, -1))
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.50 GiB (GPU 0; 15.99 GiB total capacity; 13.59 GiB already allocated; 0 bytes free; 14.21 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

Maybe, you can change the batch_size

[Hansen7777777]

but,when i change batch_size = 128, also have this question, maybe it never changes batch_size successful?

[520jz]

but,when i change batch_size = 128, also have this question, maybe it never changes batch_size successful?

change batch_size=64

[Terry10086]

Has anyone successfully run the code? If yes, when batchsize is 64(or more), how much video memory is occupied?
Looking forward to your reply~