accelbrainbase.noiseabledata package

Submodules

accelbrainbase.noiseabledata.cutout_noise module

class accelbrainbase.noiseabledata.cutout_noise.CutoutNoise(height=10, width=10)

Bases: accelbrainbase.noiseable_data.NoiseableData

Gauss noise function.

References

  • DeVries, T., & Taylor, G. W. (2017). Improved regularization of convolutional neural networks with cutout. arXiv preprint arXiv:1708.04552.
noise(arr)

Noise.

Parameters:arr – Tensor (4 or 5 rank).
Returns:Tensor.

accelbrainbase.noiseabledata.gauss_noise module

class accelbrainbase.noiseabledata.gauss_noise.GaussNoise(mu=0.0, sigma=1.0)

Bases: accelbrainbase.noiseable_data.NoiseableData

Gauss noise function.

noise(arr)

Noise.

Parameters:
  • F – mx.ndarray or mx.symbol.
  • arr – mx.nd.array or mx.sym.array.
Returns:

mx.nd.array or mx.sym.array.

Module contents