AttributeError: 'NoneType' object has no attribute 'split'

[donglin-wang]

❓ Questions and Help

I have installed the repo following the instruction on INSTALL.md. When I was following this instruction for installing apex, I used "python setup.py install" instead of "python setup.py install --cuda_ext --cpp_ext", as suggested in this issue. When I try to run the demo, I get the following error

(maskrcnn_benchmark) ParisdeMacBook-Air:demo paris$ python webcam.py --min-image-size 300 MODEL.DEVICE cpu
Traceback (most recent call last):
File "webcam.py", line 6, in
from predictor import COCODemo
File "/Users/paris/maskrcnn-benchmark/demo/predictor.py", line 6, in
from maskrcnn_benchmark.modeling.detector import build_detection_model
File "/Users/paris/maskrcnn-benchmark/maskrcnn_benchmark/modeling/detector/init.py", line 2, in
from .detectors import build_detection_model
File "/Users/paris/maskrcnn-benchmark/maskrcnn_benchmark/modeling/detector/detectors.py", line 2, in
from .generalized_rcnn import GeneralizedRCNN
File "/Users/paris/maskrcnn-benchmark/maskrcnn_benchmark/modeling/detector/generalized_rcnn.py", line 11, in
from ..backbone import build_backbone
File "/Users/paris/maskrcnn-benchmark/maskrcnn_benchmark/modeling/backbone/init.py", line 2, in
from .backbone import build_backbone
File "/Users/paris/maskrcnn-benchmark/maskrcnn_benchmark/modeling/backbone/backbone.py", line 7, in
from maskrcnn_benchmark.modeling.make_layers import conv_with_kaiming_uniform
File "/Users/paris/maskrcnn-benchmark/maskrcnn_benchmark/modeling/make_layers.py", line 10, in
from maskrcnn_benchmark.layers import Conv2d
File "/Users/paris/maskrcnn-benchmark/maskrcnn_benchmark/layers/init.py", line 10, in
from .nms import nms
File "/Users/paris/maskrcnn-benchmark/maskrcnn_benchmark/layers/nms.py", line 5, in
from apex import amp
File "", line 983, in _find_and_load
File "", line 967, in _find_and_load_unlocked
File "", line 668, in _load_unlocked
File "", line 638, in _load_backward_compatible
File "/Users/paris/anaconda3/envs/maskrcnn_benchmark/lib/python3.7/site-packages/apex-0.1-py3.7.egg/apex/init.py", line 2, in
File "", line 983, in _find_and_load
File "", line 967, in _find_and_load_unlocked
File "", line 668, in _load_unlocked
File "", line 638, in _load_backward_compatible
File "/Users/paris/anaconda3/envs/maskrcnn_benchmark/lib/python3.7/site-packages/apex-0.1-py3.7.egg/apex/amp/init.py", line 1, in
File "", line 983, in _find_and_load
File "", line 967, in _find_and_load_unlocked
File "", line 668, in _load_unlocked
File "", line 638, in _load_backward_compatible
File "/Users/paris/anaconda3/envs/maskrcnn_benchmark/lib/python3.7/site-packages/apex-0.1-py3.7.egg/apex/amp/amp.py", line 3, in
File "", line 983, in _find_and_load
File "", line 967, in _find_and_load_unlocked
File "", line 668, in _load_unlocked
File "", line 638, in _load_backward_compatible
File "/Users/paris/anaconda3/envs/maskrcnn_benchmark/lib/python3.7/site-packages/apex-0.1-py3.7.egg/apex/amp/lists/torch_overrides.py", line 69, in
File "/Users/paris/anaconda3/envs/maskrcnn_benchmark/lib/python3.7/site-packages/apex-0.1-py3.7.egg/apex/amp/utils.py", line 9, in get_cuda_version
AttributeError: 'NoneType' object has no attribute 'split'
Can anyone help me figure this out?

[RhodaLiu17]

I have the same problem as you. But then I realize that I download CPU version of pytorch-nightly, so it can't find my GPU.
Then I download GPU version of pytorch-nightly again at this website
https://anaconda.org/pytorch
Because the pip way is too slow for me, so I download .tar.bz2. Notice NOT to choose CPU!
And now it works. I hope this will be helpful!

[14211019]

I also faced this problem too,and I find the apex/amp/lists/torch._overrides.py and its 69 row is to get a GPU version ,I # it and this problem has been solved. Hope helpful for you.

[aashish-0393]

Hi.
in case of inference on CPU on machine, getting below error :

from predictor import COCODemo

error:
AttributeError: 'NoneType' object has no attribute 'split'

[xkyi]

comment:

# if version_num < 1.1:
    # FP32_FUNCS.append('mean')


# if utils.get_cuda_version() >= (9, 1, 0):
    # FP16_FUNCS.extend(_bmms)
# else:
    # FP32_FUNCS.extend(_bmms)

[BystanderJ]

I also faced this problem too,and I # (.apex/amp/lists/torch._overrides.py) 77-80 rows; and # (.apex/amp/utils.py) 8-9 rows. Then the problem was sloved, hope it's helpful for you.