tensorflow / tensorflow/probability

FFJORD: Calculating for multiple final timesteps

Open
#972 5 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Jupyter Notebook
Stars
4.4k
Forks
1.1k
PR merge metrics
No merged PRs in 30d

Description

In reference to the FFJORD bijector, I can see in its argument final_time, it only supports a single time-step (i.e. scalar).

Now, what I want is to have it calculate for multiple time-steps (i.e. say given an initial_state=x0 and initial_time = 0.0 and final_times=[1.0,2.0,3.0], then it should return calculated values for these 3 time steps). For reference, the original implementation by the authors of FFJORD paper in Pytorch supports calculating for multiple time steps: https://github.com/rtqichen/ffjord

Is there a way to implement the same in Tensorflow?

TIA

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with the FFJORD bijector and its final_time argument, then compare the linked authors' PyTorch implementation. Define the expected behavior for initial_state=x0, initial_time=0.0, and final_times=[1.0,2.0,3.0]; done means TensorFlow Probability can return values at all requested times while preserving existing scalar behavior.

Written by the indexing model from the issue text.

Assessment

Tech stack
pytorch, tensorflow
Domain
machine-learning
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.