andrewgordonwilson/bayesgan

Issues

• SGD +momentum + noise "=" SGHMC

  #18 opened 4 years ago by emiled16
  0

• d_losses is [None] while running ./run_bgan.py --data_path datasets --dataset mnist --num_mcmc 2 --out_dir ./results/ --train_iter 75000 --save_samples --n_save 100

  #17 opened 4 years ago by khorshidisamira
  0

• Missing bayesian_gan_hmc script?

  #16 opened 5 years ago by yanntrividic
  0

• The testing accuracy is not based on testing set, but rather manipulated result

  #15 opened 5 years ago by WayneDW
  0

• run_bgan does not actually switch to use specified optimizer

  #11 opened 6 years ago by julie-jiang
  1

• Why multiply d_loss_real by num_gen

  #14 opened 5 years ago by Xiaohui9607
  0

• Multiple discriminator MC samples

  #7 opened 7 years ago by ysaibhargav
  3

• Question regarding train and test parameters

  #13 opened 6 years ago by enochkan
  1

• Is bgan optimizing standard deviations?

  #12 opened 6 years ago by julie-jiang
  0

• Add WGANGP to comparison

  #4 opened 6 years ago by htt210
  1

• Custom data

  #1 opened 6 years ago by usptact
  1

• Different batch_norm for different gen_generation

  #6 opened 6 years ago by chunhanl
  4

• ValueError: Variable discriminator/d_bn1/moving_mean already exists, disallowed. Did you mean to set reuse=True or reuse=tf.AUTO_REUSE in VarScope?

  #9 opened 6 years ago by howardgriffin
  4

• Type error in semisupervised code

  #10 opened 6 years ago by whilo
  4

• Error when run bayesian_gan_hmc.py under tensorflow 1.3.0

  #5 opened 7 years ago by githubjqh
  3

• Mode collapse is a serious problem in Bayesian GAN

  #8 opened 7 years ago by htt210
  2

• 'BDCGAN' object has no attribute 'd_optim_semi_adam'

  #3 opened 7 years ago by jlandesman
  2