tensorflow / tensorflow/probability
Abstract tfp.layers' function arguments to enable safer serialization
Open
@jvdillon is already working on this.
Since Jan 24, 2019.
keras
layers
- Dominant language
- Jupyter Notebook
- Stars
- 4.4k
- Forks
- 1.1k
- PR merge metrics
- No merged PRs in 30d
Description
TFP layers use function arguments for kernel_posterior_fn, kernel_posterior_tensor_fn, kernel_prior_fn, and kernel_divergence_fn (same for bias). Following discussion with fchollet@, wicke@, ashankar@, serializing Python functions is unsafe and brittle. It is recommended we take the approach of regularizer/initializer/activation, which acts in theory like functions but are built-in and more cleanly serializable.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Assessment
This issue has not been assessed yet.