kamenbliznashki / generative_models Goto Github PK

View Code? Open in Web Editor NEW

90.0 90.0 18.0 5.47 MB

Pytorch implementations of generative models: VQVAE2, AIR, DRAW, InfoGAN, DCGAN, SSVAE

Python 100.00%

deep-learning draw infogan semi-supervised-learning

generative_models's People

Contributors

Stargazers

Watchers

Forkers

qin-folks lahiruts satyam-cyc shahuzi batikim09 smksyj njuhaozhang warvito uestcwangxiao jkartzman ahatamiz lhz1029 pkulwj1994 jbris amrakibulhasan dongso ashitaka2

generative_models's Issues

Training

Thanks for the implementation. Few questions to the training:

Does the training on other RGB dataset like COCO require any other changes besides hyperparameters of the priors?
When it comes to conditioning on class labels / captions, I'm not quite sure about the y=None in forward of priors. Does this need to be changed to refer to the one-hot encoded labels / captions?
Thanks!

GAN pre-trained weights for CheXpert

Hello and thank you for this amazing work ! I was wondering if you had shared the pre-trained weights on CheXpert for some the generative models presented in the repo (VQ-VAE, DCGAN) ?
Thank your very much for your kind help :-)

Pretrained weights and training time VQ-VAE2

Hi!
First of all congrats for the nice work.
I am wondering how long it took to train the vq-vae 2 model and the priors, and if you have the pre-trained weights from CheXpert dataset.

Best,

Recommend Projects