mindspore.nn.probability¶
Probability.
The high-level components used to construct the probabilistic network.
Bijectors¶
Bijecotr class. |
|
Exponential Bijector. |
|
GumbelCDF Bijector. |
|
Invert Bijector. |
|
Power Bijector. |
|
Scalar Affine Bijector. |
|
Softplus Bijector. |
Bayesian Layers¶
Convolutional variational layers with Reparameterization. |
|
Dense variational layers with Local Reparameterization. |
|
Dense variational layers with Reparameterization. |
Prior and Posterior Distributions¶
Build Normal distributions with trainable parameters. |
|
To initialize a normal distribution of mean 0 and standard deviation 0.1. |
Bayesian Wrapper Functions¶
Generate a suitable WithLossCell for BNN to wrap the bayesian network with loss function. |
Distributions¶
Bernoulli Distribution. |
|
Beta distribution. |
|
Create a categorical distribution parameterized by event probabilities. |
|
Cauchy distribution. |
|
Base class for all mathematical distributions. |
|
Example class: Exponential Distribution. |
|
Gamma distribution. |
|
Geometric Distribution. |
|
Gumbel distribution. |
|
Logistic distribution. |
|
LogNormal distribution. |
|
Normal distribution. |
|
Poisson Distribution. |
|
|
Transformed Distribution. |
Example class: Uniform Distribution. |
Deep Probability Networks¶
Conditional Variational Auto-Encoder (CVAE). |
|
Variational Auto-Encoder (VAE). |
Infer¶
The Evidence Lower Bound (ELBO). |
|
Stochastic Variational Inference(SVI). |
ToolBox¶
Toolbox for Uncertainty Evaluation. |
|
Toolbox for anomaly detection by using VAE. |
Model Transformer¶
Transform Deep Neural Network (DNN) model to Bayesian Neural Network (BNN) model. |