accelbrainbase.samplabledata.truesampler package

Submodules

accelbrainbase.samplabledata.truesampler.conditional_true_sampler module

class accelbrainbase.samplabledata.truesampler.conditional_true_sampler.ConditionalTrueSampler

Bases: accelbrainbase.samplabledata.true_sampler.TrueSampler

The class to draw true samples from distributions.

You should use this class when you want to build the Condtional GAN.

References

  • Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., … & Bengio, Y. (2014). Generative adversarial nets. In Advances in neural information processing systems (pp. 2672-2680).
  • Mirza, M., & Osindero, S. (2014). Conditional generative adversarial nets. arXiv preprint arXiv:1411.1784.
conditonal_dim

getter

draw()

Draws samples from the true distribution.

Returns:nd.ndarray of samples.
get_conditonal_dim()

getter

set_conditonal_dim(value)

setter

accelbrainbase.samplabledata.truesampler.labeled_true_sampler module

class accelbrainbase.samplabledata.truesampler.labeled_true_sampler.LabeledTrueSampler

Bases: accelbrainbase.samplabledata.true_sampler.TrueSampler

The class to draw true labeled samples from distributions.

References

  • Bousmalis, K., Silberman, N., Dohan, D., Erhan, D., & Krishnan, D. (2017). Unsupervised pixel-level domain adaptation with generative adversarial networks. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 3722-3731).
  • Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., … & Bengio, Y. (2014). Generative adversarial nets. In Advances in neural information processing systems (pp. 2672-2680).
  • Mirza, M., & Osindero, S. (2014). Conditional generative adversarial nets. arXiv preprint arXiv:1411.1784.
draw()

Draw samples from distribtions.

Returns:Tuple of `mx.nd.array`s.
get_labeled_image_iterator()

getter

labeled_image_iterator

getter

set_labeled_image_iterator(value)

setter

accelbrainbase.samplabledata.truesampler.normal_true_sampler module

class accelbrainbase.samplabledata.truesampler.normal_true_sampler.NormalTrueSampler(loc=0.0, scale=1.0, batch_size=40, seq_len=0, channel=3, height=96, width=96, ctx=gpu(0))

Bases: accelbrainbase.samplabledata.true_sampler.TrueSampler

The class to draw fake samples from Gaussian distributions.

References

  • Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., … & Bengio, Y. (2014). Generative adversarial nets. In Advances in neural information processing systems (pp. 2672-2680).
  • Makhzani, A., Shlens, J., Jaitly, N., Goodfellow, I., & Frey, B. (2015). Adversarial autoencoders. arXiv preprint arXiv:1511.05644.
draw()

Draw samples from distribtions.

Returns:Tuple of `mx.nd.array`s.

Module contents