Traceback (most recent call last):
File "/home/xxl/DMM/train_dmm.py", line 410, in <module>
main_function(args.experiment_directory, args.continue_from)
File "/home/xxl/DMM/train_dmm.py", line 336, in main_function
train_loss, losses_log = dmm_net(latent_vecs, sdf_data, is_on_surf, normal, centers_tensor, indices)
File "/home/xxl/anaconda3/envs/dmm/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
return forward_call(*input, **kwargs)
File "/home/xxl/DMM/networks/dmm_net.py", line 151, in forward
losses = component_sdf_normal_loss(component_sdf, coords, is_on_surf, normal,
File "/home/xxl/DMM/networks/loss.py", line 9, in component_sdf_normal_loss
gradient = compute_gradient(pred_sdf, xyz)
File "/home/xxl/DMM/utils/math.py", line 16, in compute_gradient
grad = torch.autograd.grad(y, [x], grad_outputs=grad_outputs, create_graph=True)[0]
File "/home/xxl/anaconda3/envs/dmm/lib/python3.9/site-packages/torch/autograd/__init__.py", line 276, in grad
return Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass
RuntimeError: CUDA out of memory. Tried to allocate 24.00 MiB (GPU 0; 23.70 GiB total capacity; 22.67 GiB already allocated; 6.88 MiB free; 22.70 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
my pytorch version is torch.1.12.0+cu116