8000 [associative_scan] Autograd for additional inputs by bohnstingl · Pull Request #153317 · pytorch/pytorch · GitHub
[go: up one dir, main page]

Skip to content

[associative_scan] Autograd for additional inputs #153317

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
wants to merge 34 commits into
base: main
Choose a base branch
from
Draft
Changes from 1 commit
Commits
Show all changes
34 commits
Select commit Hold shift + click to select a range
9c49a36
WIP: Associative_scan Autograd
bohnstingl Nov 6, 2024
100c598
Working implementation of Autograd
bohnstingl Nov 6, 2024
0e7c8d5
Added partial gradient tests
8000 bohnstingl Nov 7, 2024
67d62fb
Separated out the partial gradient functionality to a separate PR
bohnstingl Nov 7, 2024
9e5e1f9
Merge branch 'main' of github.com:pytorch/pytorch into associative_sc…
bohnstingl Jan 18, 2025
6b41565
Working version uncleaned
bohnstingl Mar 8, 2025
d68b31b
Almost all tests pass
bohnstingl Mar 8, 2025
9de0caf
Merge branch 'main' of github.com:pytorch/pytorch into associative_sc…
bohnstingl Mar 8, 2025
653eab0
First working implementation of simplified autograd for combine_mode=…
bohnstingl Mar 9, 2025
ac1a12b
Merge branch 'associative_scan_74' of github.com:bohnstingl/pytorch i…
bohnstingl Mar 24, 2025
d51b16c
Merge branch 'main' of github.com:pytorch/pytorch into associative_sc…
bohnstingl Mar 24, 2025
dbfd544
WIP: No additional_input support yet
bohnstingl Mar 24, 2025
64a8b29
Updates and cosmetic fixes
bohnstingl Mar 25, 2025
08b7251
iFixed problem with python<3.11
bohnstingl Mar 25, 2025
f378827
Merge branch 'main' of github.com:pytorch/pytorch into associative_sc…
bohnstingl Mar 26, 2025
b12777e
Fixed issue with adding tuple and list
bohnstingl Mar 26, 2025
6335022
Merge branch 'main' of github.com:pytorch/pytorch into associative_sc…
bohnstingl Mar 27, 2025
89ea84c
Prepared backward of associative_scan for additional_inputs
bohnstingl Mar 27, 2025
c8bbdf4
Cleaned-up code
bohnstingl Mar 27, 2025
1c9fa70
Merge branch 'main' of github.com:pytorch/pytorch into associative_sc…
bohnstingl Mar 27, 2025
6d8353b
Merge branch 'main' of github.com:pytorch/pytorch into associative_sc…
bohnstingl Apr 2, 2025
3f81639
Rework to improve readability and unify shared function with scan
bohnstingl Apr 2, 2025
62770d5
Removed irrelevant testcases, improved readability, extended document…
bohnstingl Apr 3, 2025
6761714
Restructured documentation and synced with associative_scan
bohnstingl Apr 8, 2025
b73c553
Merge branch 'main' of github.com:pytorch/pytorch into associative_sc…
bohnstingl Apr 8, 2025
ea9568d
Merge branch 'main' of github.com:pytorch/pytorch into associative_sc…
bohnstingl Apr 15, 2025
0786e6f
Merge branch 'main' of github.com:pytorch/pytorch into associative_sc…
bohnstingl Apr 24, 2025
c628dd8
Merge branch 'main' of github.com:pytorch/pytorch into associative_sc…
bohnstingl May 8, 2025
e7d63b6
Skipping autograd test for associative_scan with lifted arguments
bohnstingl May 8, 2025
95d0e42
Reworked documentation
bohnstingl May 8, 2025
3707c0d
Factored failing test out into separate test
bohnstingl May 9, 2025
dc03c1c
Merge branch 'associative_scan_74' of github.com:bohnstingl/pytorch i…
bohnstingl May 9, 2025
932f3d8
Fixed issues with additional_input imlementation
bohnstingl May 10, 2025
2edd60d
Removed resolved TODO
bohnstingl May 10, 2025
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading 8000
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Removed resolved TODO
  • Loading branch information
bohnstingl committed May 10, 2025
commit 2edd60d70210cef457d9803ad1969c8b8d3a1b67
1 change: 0 additions & 1 deletion torch/_higher_order_ops/associative_scan.py
Original file line number Diff line number Diff line change
Expand Up @@ -174,7 +174,6 @@ def add(x: torch.Tensor, y: torch.Tensor):

"""
# TODO: Support lifted arguments in inductor for associative_scan
# TODO: Support autograd for cases with lifted arguments for combine_mode=pointwise

# The reason we flatten xs before calling into dynamo is that
# we want to create a consistent input ordering for combine_fn
Expand Down
Loading
0