U-Net model for lung lesion segmentation model does not run using colab

[aficionadoai]

U-Net model for lung lesion segmentation model does not run using colab

[wyli]

for your reference the expected training outputs would be something like:

Python version: 3.6.10 |Anaconda, Inc.| (default, May  8 2020, 02:54:21)  [GCC 7.3.0]
OS version: Linux (4.15.0-50-generic)
Numpy version: 1.19.1
Pytorch version: 1.7.0a0+8deb4fe
MONAI flags: HAS_EXT = False, USE_COMPILED = False

Optional dependencies:
Pytorch Ignite version: 0.4.2
Nibabel version: 3.2.0
scikit-image version: 0.15.0
Pillow version: 8.0.1
Tensorboard version: 1.15.0+nv
gdown version: 3.12.2
TorchVision version: 0.8.0a0
ITK version: 5.1.1
tqdm version: 4.51.0

For details about installing the optional dependencies, please visit:
    https://docs.monai.io/en/latest/installation.html#installing-the-recommended-dependencies

INFO:root:training: image/label (199) folder: COVID-19-20_v2/Train
INFO:root:training: train 160 val 39, folder: COVID-19-20_v2/Train
INFO:root:batch size 2
BasicUNet features: (32, 32, 64, 128, 256, 32).
INFO:root:epochs 500, lr 0.0001, momentum 0.95
INFO:ignite.engine.engine.SupervisedTrainer:Engine run resuming from iteration 0, epoch 0 until 500 epochs
INFO:ignite.engine.engine.SupervisedTrainer:Epoch: 1/500, Iter: 1/80 -- train_loss: 1.5370 
INFO:ignite.engine.engine.SupervisedTrainer:Epoch: 1/500, Iter: 2/80 -- train_loss: 1.5101 
INFO:ignite.engine.engine.SupervisedTrainer:Epoch: 1/500, Iter: 3/80 -- train_loss: 1.4932 
...

[aficionadoai]

@wyli

Thank you! Right now, I trying to figure out why I am getting ^C right after BasicUNet features: (32, 32, 64, 128, 256, 32).. That is the part that I'm trying to figure out.

[wyli]

@wyli

Thank you! Right now, I trying to figure out why I am getting ^C right after BasicUNet features: (32, 32, 64, 128, 256, 32).. That is the part that I'm trying to figure out.

I haven't tried it with a colab instance, but perhaps you need to put

!pip install "git+https://github.com/Project-MONAI/MONAI#egg=monai[nibabel,ignite,tqdm]"

as the very first cell

[aficionadoai]

@wyli

I'm testing it out now to see if that works.

[aficionadoai]

Using it on the first cell in Cola doesn't work

[aficionadoai]

@wyli

Tentatively I think it may because I'm running into out of memory in Colab.

google colab setting a '^C' in the proccess

Colab finishes with a ^C

[wyli]

@wyli

Tentatively I think it may because I'm running into out of memory in Colab.

google colab setting a '^C' in the proccess

Colab finishes with a ^C

ok if that's the case, perhaps you could reduce the number of features for the network features=(32, 32, 64, 128, 256, 32),
btw the full training will take more than 24 hours, I'm not sure whether the colab instance supports this type of long session

[aficionadoai]

@wyli I was able to get it working with colab. You have to increase the RAM.