probability

probability copied to clipboard

Unexpected Symbolic tensor in Tensorflow Probability tensor_coercible object (mixture layer)

1

Complex interpretation of Tensorflow Probability tensor_coercible object I have a tensorflow model (keras sequential) that ends with a Tensorflow Probability (TFP) mixture layer. My goal is to fit this network...

jsaladich

Addition of "location" type parameter in the Gamma distribution

2

Adding a `loc` type parameter to the Gamma distribution like [scipy has](https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.gamma.html) would be beneficial. It is relatively common in physcs for people to use the Gamma distribution and data...

ikrommyd

NotImplementedError: Cannot convert a symbolic Tensor (truediv_4577:0) to a numpy array.

1

Hi, I am trying to train variational autoencoder with reparamaterization trick. In similar implementation on colab, I got error as **_SymbolicException: Inputs to eager execution function cannot be Keras symbolic...

DeepLearner7

mlx backend

1

Are there any plans to include [ml-explore/mlx](https://github.com/ml-explore/mlx) as `tfp` backend? It would be great for mac users running statistical computations locally. If there are no such plans, I would be...

krzysztofrusek

If I try to `jit` the `log_prob` method of the `PoissonLogNormalQuadratureCompound` ```python from jax import jit import tensorflow_probability.substrates.jax.distributions as tfd jit(tfd.PoissonLogNormalQuadratureCompound(0.0, 1.0).log_prob)(1.) ``` I get the following error: ``` TypeError:...

GianmarcoCallegher

There seems to be a bug in the model specification. in ```python tfd.MultivariateNormalDiag( loc=tf.zeros(num_students), scale_diag=self._stddev_students * tf.ones(num_students)), tfd.MultivariateNormalDiag( loc=tf.zeros(num_instructors), scale_diag=self._stddev_instructors * tf.ones(num_instructors)), tfd.MultivariateNormalDiag( loc=tf.zeros(num_departments), scale_diag=self._stddev_departments * tf.ones(num_departments)), ``` it seems...

rasoolianbehnam

How can I write a piecewise distribution? I want a distribution that consists of 3 distributions, each defined on it's range: ``` { Gamma(a-, b-)(-x), when x < 0 y...

Strategy24

The initial state for filter step is wrong, and is p(x0), when it should be p(x1|x0), i.e. kalman transition should already be applied to the initial state. This [line](https://github.com/tensorflow/probability/blob/main/tensorflow_probability/python/distributions/linear_gaussian_ssm.py#L985). Since...

Joshuaalbert

When using the JAX substrate doing optimisation of the parameters of `LinearGaussianStateSpaceModel` when I JIT compile the update step of optimisation I get the below at this line: https://github.com/tensorflow/probability/blob/v0.23.0/tensorflow_probability/python/internal/auto_composite_tensor.py#L795 ```...

Joshuaalbert

TruncatedCauchy quantile gives NaN for some parameter combinations. Similar to #1788 Numerical stability should be reinforced or it severely limits to usefulness of the distribution. # MVCE ```python import jax...

Joshuaalbert