Version 2.4.1

• @stnkl fixed a bug in probflow.utils.ops.rand_rademacher (it was previously generating all 0s 😬)
• Fixed some bugs in tests relating to the newest version of PyTorch.

Version 2.4.0

• Allow EarlyStopping to just take a MonitorMetric or MonitorELBO directly
• Remove expected_calibration_error, and replace with a more general calibration_metric, which can compute any of several different calibration metrics (like mean squared calibration error, mean absolute calibration error, miscalibration area, etc).
• Add sharpness and dispersion_metric methods to ContinuousModel (secondary uncertainty estimate metrics)
• Write the callbacks user guide section
• Some minor callbacks improvements
• Add CenteredParameter, which creates a vector of parameters which is constrained to have a mean of 0 (or a matrix whose rows and/or columns are constrained to have means of 0)

Version 2.3.0

• Add batch_norm keyword argument to DenseNetwork, which can be either True (use batch normalization between layers) or False (do not use batch normalization between layers, the default)
• Add batch_norm_loc keyword argument to DenseNetwork, which can be either 'after' (apply batch normalization after each layer's activation function) or 'before' (apply batch normalization before each layer's activation function)

Version 2.2.1

• Implement probflow.distributions.Mixture for PyTorch.

Version 2.2.0

• Add the n_mc kwarg to probflow.Model.fit, which sets the number of monte carlo samples which are taken per batch. This performs parameter updates using average of the gradients across multiple MC samples per batch. It's slower with more samples, but leads to more stable fitting.

Version 2.1.2

• Add support for flipout with the PyTorch backend.
• Add randn, rand_rademacher, and shape backend-independent ops
• Update deprecated tfp.python.math.random_rademacher in favor of tfp.random.rademacher when possible

Version 2.1.1

• Add KL divergence between Discrete distribution and other continuous distributions for PyTorch (fitting models with Deterministic parameters previously wasn't working for PyTorch)
• Refactor tests - they're nice and clean now 😊
• Add autoflake to dev stack

Version 2.1.0

• Add a probabilistic keyword argument to Dense, DenseRegression, and Embedding modules.
• Update MonitorMetric and MonitorELBO to also track walltime
• Add calibration methods to ContinuousModel: calibration_curve, calibration_curve_plot, and expected_calibration_error
• Add support for MultivariateNormalParameter for PyTorch (by implementing probflow.utils.ops.log_cholesky_transform for pytorch)
• Add a Neural linear example which uses most of these new features

Version 2.0.0

• Fixed pytorch-only import
• Added testing matrix for python versions and pytorch/tensorflow

2.0.0.a3

• Uses TensorFlow graph via tf.function - faster fitting!
• Model saving and loading (using cloudpickle)
• Fix a plotting error caused by new version of numpy/pandas
• Update TensorFlow and TensorFlow Probability version dependencies
• Add some docs and fix some other docs