csgan
csgan copied to clipboard
po0ya
Task-Aware Compressed Sensing Using Generative Adversarial Networks (published in AAAI18)
https://github.com/po0ya/csgan/blob/f1af3863d621d99c93fa94ce7f458e3cffdfec21/utils/csgan_utils.py#L657 Should change "FLAGS.n_input" to "FLAGS.n_inputs", because define "n_inputs" instead of "n_input" in https://github.com/po0ya/csgan/blob/f1af3863d621d99c93fa94ce7f458e3cffdfec21/main.py#L53
It takes about two days to train a model on the GPU. Is this normal? How can I shorten the training time? Thank you for your help.
D:\TFCode\GAN\csgan-master>py -3 main.py --cfg experiments/cfgs/csgan/dcgan_mnist.yml --generate_A Traceback (most recent call last): File "main.py", line 148, in cfg = load_config(args.cfg) File "D:\TFCode\GAN\csgan-master\utils\config.py", line 113, in load_config if hasattr(flags, k.lower()): File "C:\Python35\lib\site-packages\tensorflow\python\platform\flags.py",...
When i running the command of "./experiments/scripts/diff_measurements_base.sh ./experiments/cfgs/csgan/csgan_mnist.yml 10,20,100",Some information "+ for PID in $RUN_PIDS + wait 2806" is output on the screen and stays here.But my gpu is not...
csgan/utils/csgan_utils.py line. 541 `dataset.inv_transform(x_hats[i], type=FLAGS.input_transform_type)` `type` should be replaced by `int_type`, as the definition of `inv_transform` is `inv_transform(self, transformed_images, int_type=None)` .
Hi I am getting the following error: > ImportError: cannot import name 'NoneType' Upon a quick google search: [https://stackoverflow.com/questions/15844714/why-am-i-getting-an-error-message-in-python-cannot-import-name-nonetype](url). Do I just comment the line?
