Pymc4
| Project Name | Stars | Downloads | Repos Using This | Packages Using This | Most Recent Commit | Total Releases | Latest Release | Open Issues | License | Language |
|---|---|---|---|---|---|---|---|---|---|---|
| Probability | 3,872 |  | 126 | 213 | 4 days ago | 43 | June 07, 2022 | 621 | apache-2.0 | Jupyter Notebook |
| Probabilistic reasoning and statistical analysis in TensorFlow | ||||||||||
| Ml Foundations | 1,705 | 5 months ago | mit | Jupyter Notebook | ||||||
| Machine Learning Foundations: Linear Algebra, Calculus, Statistics & Computer Science | ||||||||||
| Pymc4 | 708 | a year ago | 2 | January 25, 2020 | 49 | apache-2.0 | Jupyter Notebook | |||
| Experimental PyMC interface for TensorFlow Probability. Official work on this project has been discontinued. | ||||||||||
| Wtte Rnn | 580 |  | 1 | 3 years ago | 5 | June 21, 2018 | 33 | mit | Python | |
| WTTE-RNN a framework for churn and time to event prediction | ||||||||||
| Tfcausalimpact | 366 |  | 3 | 3 months ago | 20 | May 03, 2022 | 25 | apache-2.0 | Python | |
| Python Causal Impact Implementation Based on Google's R Package. Built using TensorFlow Probability. | ||||||||||
| Googlenet Inception | 232 | 3 years ago | 3 | mit | Python | |||||
| TensorFlow implementation of GoogLeNet and Inception for image classification. | ||||||||||
| Deeplearningwithtf2.0 | 227 | 3 years ago | Jupyter Notebook | |||||||
| Practical Exercises in TensorFlow 2.0 for Ian Goodfellows Deep Learning Book | ||||||||||
| Rethinking Tensorflow Probability | 207 | a year ago | apache-2.0 | Jupyter Notebook | ||||||
| Statistical Rethinking (2nd Ed) with Tensorflow Probability | ||||||||||
| Fenchel Young Losses | 175 | 5 months ago | Python | |||||||
| Probabilistic classification in PyTorch/TensorFlow/scikit-learn with Fenchel-Young losses | ||||||||||
| C3d Keras | 159 | 6 years ago | 5 | other | Python | |||||
| C3D for Keras + TensorFlow | ||||||||||
NOTICE: Official development of this project has ceased, and it is no longer intended to become the next major version of PyMC. Ongoing development will continue on the main PyMC repository.
See the announcement for more details on the future of PyMC and Theano.
High-level interface to TensorFlow Probability. Do not use for anything serious.
What works?
- Build most models you could build with PyMC3
- Sample using NUTS, all in TF, fully vectorized across chains (multiple chains basically become free)
- Automatic transforms of model to the real line
- Prior and posterior predictive sampling
- Deterministic variables
- Trace that can be passed to ArviZ
However, expect things to break or change without warning.
See here for an example: https://github.com/pymc-devs/pymc4/blob/master/notebooks/radon_hierarchical.ipynb See here for the design document: https://github.com/pymc-devs/pymc4/blob/master/notebooks/pymc4_design_guide.ipynb
Develop
One easy way of developing on PyMC4 is to take advantage of the development containers! Using pre-built development environments allows you to develop on PyMC4 without needing to set up locally.
To use the dev containers, you will need to have Docker and VSCode running locally on your machine,
and will need the VSCode Remote extension (ms-vscode-remote.vscode-remote-extensionpack).
Once you have done that, to develop on PyMC4, on GitHub:
- Make a fork of the repository
- Create a new branch inside your fork
- Copy the branch URL
Now, in VSCode:
- In the command palette, search for "Remote-Containers: Open Repository in Container...".
- Paste in the branch URL
- If prompted, create it in a "Unique Volume".
Happy hacking away! Because the repo will be cloned into an ephemeral repo, don't forget to commit your changes and push them to your branch! Then follow the usual pull request workflow back into PyMC4.
We hope you enjoy the time saved on setting up your development environment!
