I think that data["img"] should be a list of 2 element with sizes [2, 3, 50, 50].
After checking collate implementation I saw that bare tensors are just being stacked. So probably we should not be using collate directly after a pipeline transformation? Is there another way for doing this?
I am willing to collaborate if needed.
Thanks :)
Environment
Previous command is not working for me
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/home/mmendez/.cache/pypoetry/virtualenvs/ai-homography-estimation-MIZJYHGs-py3.8/lib/python3.8/site-packages/mmcls/utils/collect_env.py", line 10, in collect_env
env_info = collect_base_env()
File "/home/mmendez/.cache/pypoetry/virtualenvs/ai-homography-estimation-MIZJYHGs-py3.8/lib/python3.8/site-packages/mmcv/utils/env.py", line 112, in collect_env
from mmcv.ops import get_compiler_version, get_compiling_cuda_version
File "/home/mmendez/.cache/pypoetry/virtualenvs/ai-homography-estimation-MIZJYHGs-py3.8/lib/python3.8/site-packages/mmcv/ops/__init__.py", line 2, in <module>
from .active_rotated_filter import active_rotated_filter
File "/home/mmendez/.cache/pypoetry/virtualenvs/ai-homography-estimation-MIZJYHGs-py3.8/lib/python3.8/site-packages/mmcv/ops/active_rotated_filter.py", line 10, in <module>
ext_module = ext_loader.load_ext(
File "/home/mmendez/.cache/pypoetry/virtualenvs/ai-homography-estimation-MIZJYHGs-py3.8/lib/python3.8/site-packages/mmcv/utils/ext_loader.py", line 13, in load_ext
ext = importlib.import_module('mmcv.' + name)
File "/usr/lib/python3.8/importlib/__init__.py", line 127, in import_module
return _bootstrap._gcd_import(name[level:], package, level)
ImportError: /home/mmendez/.cache/pypoetry/virtualenvs/ai-homography-estimation-MIZJYHGs-py3.8/lib/python3.8/site-packages/
mmcv/_ext.cpython-38-x86_64-linux-gnu.so: undefined symbol: _ZN2at5sliceERKNS_6TensorElN3c108optionalIlEES5_l
Describe the bug
I am trying to customize
inference_model
function so it can work with batches.It seems the collate is not working as expected. This would serve as a simple example:
I think that
data["img"]
should be a list of 2 element with sizes [2, 3, 50, 50].After checking
collate
implementation I saw that bare tensors are just being stacked. So probably we should not be using collate directly after a pipeline transformation? Is there another way for doing this?I am willing to collaborate if needed.
Thanks :)
Environment
Previous command is not working for me
I have run pip list instead:
Other information
No response