dustin
October 10, 2018, 10:37pm
2
Unfortunately no re:maintenance. I have my hands tied across quite a few systems and research projects. I recommend TF’s stack overflow and tfprobability@tensorflow.org .
When the NIPS camera-ready for Edward2 is released, more details should be available. In the meantime, there’s documentation here:
# Edward2
Edward2 is a probabilistic programming language in TensorFlow and Python. It
extends the TensorFlow ecosystem so that one can declare models as
probabilistic programs and manipulate a model's computation for flexible
training, latent variable inference, and predictions.
Are you upgrading from Edward? Check out the guide
[`Upgrading_from_Edward_to_Edward2.md`](https://github.com/tensorflow/probability/blob/master/tensorflow_probability/python/edward2/Upgrading_From_Edward_To_Edward2.md).
## 1. Models as Probabilistic Programs
### Random Variables
In Edward2, we use
[`RandomVariables`](https://github.com/tensorflow/probability/blob/master/tensorflow_probability/python/edward2/random_variable.py)
to specify a probabilistic model's structure.
A random variable `rv` carries a probability distribution (`rv.distribution`),
which is a TensorFlow Distribution instance governing the random variable's methods
such as `log_prob` and `sample`.
show original
1 Like