Revert prefactor m in autograd documentation #3221
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/tutorials/3221
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit dc20fcb with merge base 82f449a ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
|
Hi @asideofcode-dev! Thank you for your pull request and welcome to our community. Action RequiredIn order to merge any pull request (code, docs, etc.), we require contributors to sign our Contributor License Agreement, and we don't seem to have one on file for you. ProcessIn order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA. Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with If you have received this in error or have any questions, please contact us at [email protected]. Thanks! |
svekars
added
the
core
|
Thank you for signing our Contributor License Agreement. We can now accept your code for this (and any) Meta Open Source project. Thanks! |
The prefactor m was incorrectly added to the vector-Jacobian product formula in the autograd tutorial. This change was based on the mistaken assumption that m-scaling was necessary to account for multiple terms in the summation. However, the chain rule and vector-Jacobian product formula already correctly aggregate contributions from all intermediate variables without requiring explicit scaling by m.
asideofcode-dev
force-pushed
the
patch-1
branch
from
a4363ef to
dc20fcb
Compare
|
@albanD No worries! Thanks for approving. I think some workflows need approval before moving forward... not 100% sure how the approval process works but I'll leave the rest in your hands :) Thanks again for the prompt approval |
|
I'll let @svekars check that part! |
Fixes #2168
Description
The prefactor m was incorrectly added to the vector-Jacobian product formula in the autograd tutorial as part of #3181. This change was based on the mistaken assumption that m-scaling was necessary to account for multiple terms in the summation. However, the chain rule and vector-Jacobian product formula already correctly aggregate contributions from all intermediate variables without requiring explicit scaling by m.
For a deeper explanation see #2168 (comment).
Checklist
cc @albanD @jbschlosser