Error while running - ValueError: axis(=1) out of bounds
circularProteus opened this issue · 2 comments
I'm having problem using the script in command line;
Here the following result:
(C:\local\Anaconda3-4.1.1-Windows-x86_64\envs\cntk-py35) C:\local\CNTK-2-0-rc3\cntk\Examples\Image\Detection\FastRCNN>python "C:\Users\Ale\Desktop\CNTK-FastRCNNDetector-master\frcnn_detector.py" --input C:\imagetest\1.jpg --model C:\local\CNTK-2-0-rc3\cntk\Examples\Image\Detection\FastRCNN\proc\Grocery_100\cntkFiles\Output\Fast-RCNN.model --output C:\output
and the final:
Selected CPU as the process wide default device.
Number of images to process: 1
Read file in path: C:\imagetest\1.jpg
C:\local\Anaconda3-4.1.1-Windows-x86_64\envs\cntk-py35\lib\site-packages\cntk\core.py:349: UserWarning: your data is of type "float64", but your input variable (uid "Input224") expects "<class 'numpy.float32'>". Please convert your data beforehand to speed up training.
(sample.dtype, var.uid, str(var.dtype)))
C:\local\Anaconda3-4.1.1-Windows-x86_64\envs\cntk-py35\lib\site-packages\cntk\core.py:349: UserWarning: your data is of type "float64", but your input variable (uid "Input225") expects "<class 'numpy.float32'>". Please convert your data beforehand to speed up training.
(sample.dtype, var.uid, str(var.dtype)))
Traceback (most recent call last):
File "C:\Users\Ale\Desktop\CNTK-FastRCNNDetector-master\frcnn_detector.py", line 361, in
rects, labels = detector.detect(img)
File "C:\Users\Ale\Desktop\CNTK-FastRCNNDetector-master\frcnn_detector.py", line 277, in detect
rois_labels_predictions = np.argmax(rois_values, axis=1)
File "C:\local\Anaconda3-4.1.1-Windows-x86_64\envs\cntk-py35\lib\site-packages\numpy\core\fromnumeric.py", line 963, in argmax
return _wrapfunc(a, 'argmax', axis=axis, out=out)
File "C:\local\Anaconda3-4.1.1-Windows-x86_64\envs\cntk-py35\lib\site-packages\numpy\core\fromnumeric.py", line 57, in _wrapfunc
return getattr(obj, method)(*args, **kwds)
ValueError: axis(=1) out of bounds
Hi @alessi89, thanks for reporting this issue. It was introduced by a change in the CNTK API.
I fixed it to use the changed API and the module seems to work now.
Let me know if you run into any more issues.
ok it works. Thanks