tensorflow / tensorflow/probability
Feature Request: log_prob_parts() for Blockwise
Nobody has claimed this yet.
- Dominant language
- Jupyter Notebook
- Stars
- 4.4k
- Forks
- 1.1k
- PR merge metrics
- No merged PRs in 30d
Description
Currently, the log_prob method for a tfp.distributions.Blockwise returns a one-dimensional tensor only, with the log likelihoods summed across the component distributions. It would be useful if there was a log_prob_parts() method as in joint_distribution_sequential, that returned a tensor of element-wise log-likelihoods, to facilitate easy masking or weighting of the likelihoods in custom loss functions.
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.
Research direction
Start with the Blockwise log_prob entry point and compare its behavior with log_prob_parts() in joint_distribution_sequential. The work is complete when Blockwise exposes element-wise component log-likelihoods that can be masked or weighted independently in custom loss functions.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- tensorflow
- Domain
- machine-learning
- Issue type
- Feature
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 38/100