Closed abhineet99 closed 4 years ago
hello,Do you sucess run this repo?
Yes I did, had to make some changes though. Might be able to help if you can pin point the issues you're facing?
@abhineet99 Thanks for yorr help!
I want to ask what environment you use?
I use conda python2.7 pytorch0.4 and torchvision 0.2.1
When I run start_pretraining.sh by this command: python runtime.py --type=pretrain_caption --dataset=flowers
I got this error:
/home/dany/text2image2textGAN-master/txt2image_dataset.py:28: ParserWarning: Falling back to the 'python' engine because the 'c' engine does not support regex separators (separators > 1 char and different from '\s+' are interpreted as regex); you can avoid this warning by specifying engine='python'.
self.image_paths_df = pd.read_table('data/images.txt', sep='\s+|\/+', header=None)
False
pretrain_caption
gan
0%| | 0/100 [00:07<?, ?it/s]
Traceback (most recent call last):
File "runtime.py", line 52, in
Traceback (most recent call last): File "runtime.py", line 55, in cycle_trainer.train(args.cls)
NameError: name 'cycle_trainer' is not defined
Do you have the code available for cycle_gan end to end training? It seems that this has been accepted as an argument but no implementation is available in trainer.py