ngessert / isic2018 Goto Github PK

hello sir,
when trying to run your code, for example:
python train.py example ISIC.example_senet_5foldcv

this error appear:

Traceback (most recent call last): File "train.py", line 38, in <module> import models File "/content/models.py", line 8, in <module> import pretrainedmodels ModuleNotFoundError: No module named 'pretrainedmodels'

Thank you for your code!
I have a question, do I need to find all the pre-training models myself?
When I run train.py,it says: No module named 'pretraindmodels'.

Hi!
Thanks for opening your code. It is really a big help for other researchers.

In you paper, you mentioned that "All images were separated based on lesion affiliation,
i.e., we made sure that images from the same lesion cannot occur both in a training and validation split."

I know this strategy do make a difference when training, but i don't know how to get the lesion affiliation.

i only have HAM10000 images so i tried example and followed everything and solve paths errors on colab but after i solved those errors i got this error after reach Evaluating on validation

['/content/isic2018/data/isic/task3/images/HAM10000/']
/usr/local/lib/python3.6/dist-packages/numpy/core/fromnumeric.py:3257: RuntimeWarning: Mean of empty slice.
out=out, **kwargs)
/usr/local/lib/python3.6/dist-packages/numpy/core/_methods.py:154: RuntimeWarning: invalid value encountered in true_divide
ret, rcount, out=ret, casting='unsafe', subok=False)
[nan nan nan nan nan nan nan]
Set sizes [10015.]
Set Names {0: '/content/isic2018/data/isic/task3/images/HAM10000/'}
Set starts [1.0000e+00 1.0016e+04]
[24306 24307 24308 ... 34318 34319 34320]
Train
(8009,)
(8005,)
(8008,)
(8006,)
(8006,)
Val
(2006,)
(2010,)
(2007,)
(2009,)
(2009,)
Positions [[112 112]
[112 237]
[112 362]
[112 488]
[187 112]
[187 237]
[187 362]
[187 488]
[262 112]
[262 237]
[262 362]
[262 488]
[338 112]
[338 237]
[338 362]
[338 488]]
Shape 1 (224, 224, 3)
Shape 2 (224, 224, 3)
Shape 3 (224, 224, 3)
Shape 4 (224, 224, 3)
Shape 5 (224, 224, 3)
Shape 6 (224, 224, 3)
Shape 7 (224, 224, 3)
Shape 8 (224, 224, 3)
Shape 9 (224, 224, 3)
Shape 10 (224, 224, 3)
Shape 11 (224, 224, 3)
Shape 12 (224, 224, 3)
Shape 13 (224, 224, 3)
Shape 14 (224, 224, 3)
Shape 15 (224, 224, 3)
Shape 16 (224, 224, 3)
Devices to use: 0,1,2,3
Evaluating on validation set during training.
cuda:0
Traceback (most recent call last):
File "/content/isic2018/train.py", line 163, in
class_weights = 1.0/np.mean(mdlParams['labels_array'][indices_ham,:],axis=0)
IndexError: index 4 is out of bounds for axis 0 with size 0