python vggprune.py about network-slimming HOT 6 CLOSED

Could you share more details about the fault? For example, the traceback of the fault. It would be more helpful for us to look into your problem.

from network-slimming.

first i run python main.py --dataset cifar10 --arch vgg --depth 19, and then python main.py -sr --s 0.0001 --dataset cifar10 --arch vgg --depth 19
i got two file, checkpoint.pth.tar and model_best.pth.tar.
run python vggprune.py, i set check.pth.tar is the model path and get :

=> loading checkpoint './logs/checkpoint.pth.tar'
Traceback (most recent call last):
File "vggprune.py", line 43, in
model.load_state_dict(checkpoint['state_dict'])
File "/usr/local/lib/python3.5/dist-packages/torch/nn/modules/module.py", line 522, in load_state_dict
.format(name))
KeyError: 'unexpected key "module.feature.0.weight" in state_dict'

from network-slimming.

I try the code myself. There is not any problem with the code itself.

I think the problem is that the format of checkpoint.pth.tar and variable model is not the same. One of them is a DataParrell object. Please make sure they conform to the same object type.

from network-slimming.

Thanks for your nice sharing!
there is a fault when i run the file of vggprune.py.
Traceback (most recent call last):
File "vggprune.py", line 142, in
m1.weight.data = m0.weight.data[idx1.tolist()].clone()
TypeError: indexing a tensor with an object of type list. The only supported types are integers, slices, numpy scalars and torch.cuda.LongTensor or torch.cuda.ByteTensor as the only argument.

from network-slimming.

@nepulhl are you using Pytorch 0.3.1? Our code is developed based on Pytorch version 0.3.1.

from network-slimming.

@Eric-mingjie Thank you,I have sloved the problem by updating Pytorch.

from network-slimming.