pymc-devs / pymc-devs/pytensor

Rewrite to remove useless transpose followed by reduction

Open
#1,006 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

graph rewriting
Dominant language
Python
Stars
644
Forks
208
Avg merge
2d 14h
Merged PRs (30d)
16

Description

Description
import pytensor
import pytensor.tensor as pt
x = pt.matrix("x")
y = pt.sum(x.T)
fn = pytensor.function([x], y)
fn.dprint()
Sum{axes=None} [id A] 1
 └─ Transpose{axes=[1, 0]} [id B] 'x.T' 0
    └─ x [id C]

If we can remove a transpose (dimshuffle) without affecting the output (sometimes by changing the reduction axes, sometimes without having to do anything), we could rewrite it away. That allows more succinct graphs and more extensive canonicalizaiton

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 by reproducing the Python example with pt.sum(x.T) and inspecting its printed graph. Trace the graph-rewrite or optimization entry point that handles reductions and transposes, then verify that the transpose is absent while the result remains equivalent for the relevant axes.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
backend
Issue type
Refactor
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
45/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.