Describe the bug
RandHistogramShift transform gives nan values when an input with all zeros is passed to it.
To Reproduce
import monai
import torch
transform = monai.transforms.RandHistogramShift(prob=1)
tensor = torch.zeros((2, 2))
print(transform(tensor))
gives the following result,
tensor([[nan, nan],
[nan, nan]])
Expected behavior
The returned tensor should contain zeros instead of nan values.
In a previous version of monai, this worked fine as it was called on a ndarray by default. In the current version, even if an ndarray is passed, the same nan behaviour is thrown. It looks like the functionality to handle a ndarray
|
if isinstance(x, np.ndarray): |
is never called because of
|
img = convert_to_tensor(img, track_meta=get_track_meta()) |
Environment
================================
Printing MONAI config...
================================
MONAI version: 1.1.0+29.gbdf5e1ec
Numpy version: 1.23.5
Pytorch version: 1.13.1+cu117
MONAI flags: HAS_EXT = False, USE_COMPILED = False, USE_META_DICT = False
MONAI rev id: bdf5e1ec15724ecf4e3634955cf34a35843a3e3e
MONAI __file__: /home/suraj/anaconda3/envs/monai-dev/lib/python3.8/site-packages/monai/__init__.py
Optional dependencies:
Pytorch Ignite version: 0.4.10
ITK version: 5.3.0
Nibabel version: 5.0.0
scikit-image version: 0.19.3
Pillow version: 9.4.0
Tensorboard version: 2.11.2
gdown version: 4.6.0
TorchVision version: 0.14.1+cu117
tqdm version: 4.64.1
lmdb version: 1.4.0
psutil version: 5.9.4
pandas version: 1.5.2
einops version: 0.6.0
transformers version: 4.21.3
mlflow version: 2.1.1
pynrrd version: 1.0.0
For details about installing the optional dependencies, please visit:
https://docs.monai.io/en/latest/installation.html#installing-the-recommended-dependencies
================================
Printing system config...
================================
System: Linux
Linux version: Ubuntu 20.04.4 LTS
Platform: Linux-5.13.0-40-generic-x86_64-with-glibc2.10
Processor: x86_64
Machine: x86_64
Python version: 3.8.15
Process name: python
Command: ['python']
Num physical CPUs: 12
Num logical CPUs: 12
Num usable CPUs: 12
CPU usage (%): [13.4, 8.7, 11.2, 10.7, 8.4, 6.8, 5.8, 5.7, 50.0, 8.4, 6.0, 27.5]
CPU freq. (MHz): 914
Load avg. in last 1, 5, 15 mins (%): [6.7, 13.4, 47.2]
Disk usage (%): 78.3
Avg. sensor temp. (Celsius): UNKNOWN for given OS
Total physical memory (GB): 250.6
Available memory (GB): 93.0
Used memory (GB): 5.7
================================
Printing GPU config...
================================
Num GPUs: 4
Has CUDA: True
CUDA version: 11.7
cuDNN enabled: True
cuDNN version: 8500
Current device: 0
Library compiled for CUDA architectures: ['sm_37', 'sm_50', 'sm_60', 'sm_70', 'sm_75', 'sm_80', 'sm_86']
GPU 0 Name: Quadro RTX 8000
GPU 0 Is integrated: False
GPU 0 Is multi GPU board: False
GPU 0 Multi processor count: 72
GPU 0 Total memory (GB): 47.5
GPU 0 CUDA capability (maj.min): 7.5
GPU 1 Name: Quadro RTX 8000
GPU 1 Is integrated: False
GPU 1 Is multi GPU board: False
GPU 1 Multi processor count: 72
GPU 1 Total memory (GB): 47.5
GPU 1 CUDA capability (maj.min): 7.5
GPU 2 Name: Quadro RTX 8000
GPU 2 Is integrated: False
GPU 2 Is multi GPU board: False
GPU 2 Multi processor count: 72
GPU 2 Total memory (GB): 47.5
GPU 2 CUDA capability (maj.min): 7.5
GPU 3 Name: Quadro RTX 8000
GPU 3 Is integrated: False
GPU 3 Is multi GPU board: False
GPU 3 Multi processor count: 72
GPU 3 Total memory (GB): 47.5
GPU 3 CUDA capability (maj.min): 7.5
Describe the bug
RandHistogramShifttransform givesnanvalues when an input with all zeros is passed to it.To Reproduce
gives the following result,
Expected behavior
The returned tensor should contain zeros instead of
nanvalues.In a previous version of monai, this worked fine as it was called on a
ndarrayby default. In the current version, even if anndarrayis passed, the samenanbehaviour is thrown. It looks like the functionality to handle andarrayMONAI/monai/transforms/intensity/array.py
Line 1433 in bdf5e1e
is never called because of
MONAI/monai/transforms/intensity/array.py
Line 1461 in bdf5e1e
Environment