probability

probability copied to clipboard

nvm

pasq-cat

Enable batch support for `windowed_mean|variance`

5

This PR makes functions `windowed_mean` and `windowed_variance` to accept indices with batch dimensions. Example: ```python x = np.array([[1, 2, 3], [1, 2, 3], [1, 2, 3]], dtype=np.float32) low_indices = [[0,...

nicolaspi

Add `tfp.math.betaincinv`

2

This PR provides an implementation for `tfp.math.betaincinv`, the inverse of the regularized incomplete beta function (`tfp.math.betainc`) with respect to `x`. It addresses the issue #1581. This implementation is based on...

leandrolcampos

The inverse of the regularized incomplete beta function

4

Hello TFP community, I’d like to work on the implementation of the inverse of the regularized incomplete beta function (`special.betaincinv` in SciPy). To do this, I plan to study the...

leandrolcampos

Feature Request: Gumbel Mixture Models

2

It is possible to construct reparameterizable mixture distributions by replacing the categorical distribution with a gumbel (relaxed categorical) distribution. The ability to use a relaxed one-hot categorical distribution in mixture...

bryorsnef

tf.math.betainc is missing gradients

6

I'm trying to fit a `NegativeBinomial`, but its `cdf` seems to be missing some gradients: ```python count = tf.compat.v1.get_variable("count", shape=()) logit = tf.compat.v1.get_variable("logit", shape=()) value = tf.compat.v1.get_variable("value", shape=()) cdf =...

jesnie

Estimate prediction uncertainty in ResNet model using MC dropout

1

is there a tutorial on how to apply TFP with popular vision models such as ResNet using MC dropout, and extract aleatoric & epistemic uncertainties. Perhaps something similar to this...

akamil-etsy

I want to build a distribution layer of multivariate students't distribution. when I using tensorflow.distribution.MultivariateStudentTLinearOperator in DistributionLambda, always have an error: `AttributeError: Exception encountered when calling layer "distribution_lambda_20" (type DistributionLambda)....

yang24feifei

probability 0.17.0 doesn't build on macOS with error ModuleNotFoundError: No module named 'absl'

1

I'm attempting to update `tensor flow_probability` for MacPorts. Version 0.17.0 fails to build on macOS with error: > ``` > File "/opt/local/var/macports/build/_opt_local_ports_python_py-tensorflow-probability/py310-tensorflow-probability/work/bazel_build/4191c7127d7c33c5af5f12aa8dc0d886/sandbox/processwrapper-sandbox/5/execroot/tensorflow_probability/bazel-out/darwin-opt-exec-2B5CBBC6/bin/tensorflow_probability/substrates/meta/rewrite.runfiles/tensorflow_probability/tensorflow_probability/substrates/meta/rewrite.py", line 20, in > from absl import app...

essandess

Fix `scale_tril` argument in Gaussian_Copula example notebook

3

Fix `scale_tril` argument, so marginals are actually uniform, as stated in the primer text. See https://github.com/tensorflow/probability/issues/1591

gmjw