🐛 Describe the bug
The function pypose.quat2unit modifies tensors in place. E.g.,
import pypose
a = pypose.randn_SE3(requires_grad=True)
b = pypose.quat2unit(a)
b.tensor().mean().backward()
raises
Traceback (most recent call last):
File "/root/test.py", line 5, in <module>
b = pypose.quat2unit(a)
File "/miniforge/envs/pypose/lib/python3.10/site-packages/pypose/lietensor/convert.py", line 855, in quat2unit
data[..., 3:7] = normalize(data[..., 3:7], p=2, dim=-1, eps=eps)
RuntimeError: a view of a leaf Variable that requires grad is being used in an in-place operation.
a potential fix is to change these lines to
if input.ltype in [SO3_type, RxSO3_type]:
quat, *scale = torch.split(data, [4, input.ltype.dimension[0] - 4], -1)
quat = normalize(quat, p=2, dim=-1, eps=eps)
data = torch.concatenate((quat, *scale), -1)
elif input.ltype in [SE3_type, Sim3_type]:
t, quat, *scale = torch.split(data, [3, 4, input.ltype.dimension[0] - 7], -1)
quat = normalize(quat, p=2, dim=-1, eps=eps)
data = torch.concatenate((t, quat, *scale), -1)Versions
PyTorch version: 2.4.1
Is debug build: False
CUDA used to build PyTorch: None
ROCM used to build PyTorch: N/A
OS: macOS 15.2 (arm64)
GCC version: Could not collect
Clang version: 16.0.0 (clang-1600.0.26.6)
CMake version: version 3.30.5
Libc version: N/A
Python version: 3.10.15 (main, Sep 9 2024, 22:43:48) [Clang 18.1.8 ] (64-bit runtime)
Python platform: macOS-15.2-arm64-arm-64bit
Is CUDA available: False
CUDA runtime version: No CUDA
GPU models and configuration: No CUDA
Nvidia driver version: No CUDA
cuDNN version: No CUDA
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True
Versions of relevant libraries:
[pip3] mypy-boto3-s3==1.35.32
[pip3] mypy-extensions==1.0.0
[pip3] numpy==1.26.4
[pip3] torch==2.4.1
[pip3] torch_scatter==2.1.2
[pip3] torch_sparse==0.6.18
[pip3] torch-tb-profiler==0.4.3
[pip3] torchmetrics==1.5.1
[pip3] torchvision==0.19.1
[pip3] torchviz==0.0.2
[conda] numpy 1.23.5 pypi_0 pypi
[conda] numpydoc 1.7.0 py310hca03da5_0
[conda] torch 2.3.1 pypi_0 pypi
[conda] torch-tb-profiler 0.4.3 pypi_0 pypi
[conda] torchvision 0.18.1 pypi_0 pypi
🐛 Describe the bug
The function
pypose.quat2unitmodifies tensors in place. E.g.,raises
a potential fix is to change these lines to
Versions