8000 pinv: forward/backward AD which is Frechet-defined in a rank-preserving neighborhood. by nikitaved · Pull Request #66092 · pytorch/pytorch · GitHub
[go: up one dir, main page]
Skip to content

pinv: forward/backward AD which is Frechet-defined in a rank-preserving neighborhood. #66092

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.

Sign up for GitHub

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

Closed
wants to merge 12 commits into from
Closed

pinv: forward/backward AD which is Frechet-defined in a rank-preserving neighborhood. #66092

Changes from 1 commit
Commits
Show all changes
12 commits
Select commit Hold shift + click to select a range
d4ddef8
pinv: forward/backward AD which is Frechet differentiable in a
nikitaved Sep 30, 2021
1ad418e
add forward AD to the tests of pinverse in OpInfo
nikitaved Oct 5, 2021
4775c45
Merge branch 'master' of https://github.com/pytorch/pytorch into niki…
nikitaved Oct 5, 2021
346113f
minor optimization
nikitaved Oct 5, 2021
23c6ccf
Merge branch 'master' of https://github.com/pytorch/pytorch into niki…
nikitaved Oct 7, 2021
b154a16
minor test update
nikitaved Oct 7, 2021
3743c7d
Merge branch 'master' of https://github.com/pytorch/pytorch into niki…
nikitaved Oct 7, 2021
2dbbecb
add derivations
nikitaved Oct 7, 2021
443ccc1
Merge branch 'master' of https://github.com/pytorch/pytorch into niki…
nikitaved Oct 7, 2021
378abc5
minor
nikitaved Oct 8, 2021
0ec1de8
resolve merge conflict
nikitaved Oct 8, 2021
2b5431e
fix flake8
nikitaved Oct 8, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
minor
  • Loading branch information
@nikitaved
nikitaved committed Oct 8, 2021
commit 378abc5c8f2c02f26f461107c4d134a19256d51c
2 changes: 1 addition & 1 deletion torch/testing/_internal/common_methods_invocations.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2121,7 +2121,7 @@ def generate_samples():
for batch, m, n in product(batches, size, size):
for k in range(min(3, min(m, n))):
# Note that by making the columns of `a` and `b` orthonormal we make sure that
# the product matrix `a @ b.t()` has condition number when restricted to its image
# the product matrix `a @ b.t()` has condition number 1 when restricted to its image
a = torch.rand(*batch, m, k, device=device, dtype=dtype).qr().Q.requires_grad_(requires_grad)
b = torch.rand(*batch, n, k, device=device, dtype=dtype).qr().Q.requires_grad_(requires_grad)
yield SampleInput(a, args=(b,))
Expand Down
0