RuntimeError: leaf variable has been moved into the graph interior

[novauto-nju]

When I reduce the BatchSize training, the program reports an error:

Traceback (most recent call last):
File "train_deep_sdf.py", line 589, in
main_function(args.experiment_directory, args.continue_from, int(args.batch_split))
File "train_deep_sdf.py", line 510, in main_function
chunk_loss.backward()
File "/home/zenghui/.conda/envs/py37/lib/python3.7/site-packages/torch/tensor.py", line 166, in backward
torch.autograd.backward(self, gradient, retain_graph, create_graph)
File "/home/zenghui/.conda/envs/py37/lib/python3.7/site-packages/torch/autograd/init.py", line 99, in backward
allow_unreachable=True) # allow_unreachable flag
RuntimeError: leaf variable has been moved into the graph interior