Hi @rampage644, i am trying your pixelcnn implementation on cifar with python train.py -g0 --levels 256 --out data/ --dataset cifar, but I am receiving assertion error AttributeError: 'MaskedConvolution2D' object has no attribute 'has_uninitialized_params'. Why is it? I think I have all the dependencies installed.
I think this is because some dependencies were updated (the code was developed almost 2 year ago). I didn't pin the version I've used as you can see.
Feel free to submit a fix.
Hi @rampage644, i am trying your pixelcnn implementation on cifar with
python train.py -g0 --levels 256 --out data/ --dataset cifar
, but I am receiving assertion errorAttributeError: 'MaskedConvolution2D' object has no attribute 'has_uninitialized_params'
. Why is it? I think I have all the dependencies installed.