some smallimageNet training questions

[jiajunhua]

Firstly, thank you for sharing your great works and sorry for my poor English.

I have a question about smallimageNet. Pictures pixels are 64*64 ,I change train_imagenet.sh the 19 line's raw_size to 64,then run train_imagenet.sh ,some mistakes happened in follows: Pre-sized data not found (./data/imagenet64); using original data: ./data/imagenet Traceback (most recent call last): File "train_gan.py", line 262, in dataset = data.Dataset(args) File "/home/hddsj/bigan-master/data.py", line 453, in init max_images=args.max_images) File "/home/hddsj/bigan-master/data.py", line 210, in imagenet_data_providers assert 0 <= num_test <= len(data['val']) AssertionError Then run resize_imageset.py some errors appears，in follows: usage: resize_imageset.py [-h] [-c] [-f FORMAT] [-q] [-j JOBS] [-r] [-o] size input_directory output_directory resize_imageset.py: error: too few arguments

How can I input these arguments ???

What does the following code in Readme mean ??? SIZE=72 # or SIZE=128 for generalized BiGAN experiments

"-j 4" uses 4 resizing processes

python resize_imageset.py -r -j 4 ${SIZE} ./data/imagenet ./data/imagenet${SIZE}

[jiajunhua]

Thank you very much! My email is [email protected] I looking forward for your reply.