dev(svd): add support for svd #70
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Collaborator
gausshj
commented
Oct 9, 2025
gausshj
commented
- Add support for svd
- Add test for svd
- Add support for svd - Add test for svd Signed-off-by: Gausshj <[email protected]>
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
This was referenced Oct 9, 2025
hzhangxyz
requested changes
Oct 9, 2025
grassmann_tensor/tensor.py
Outdated
|
|
||
| tensor = tensor.reshape((left_dim, right_dim)) | ||
|
|
||
| U, S, Vh = torch.linalg.svd(tensor.tensor, full_matrices=full_matrices) |
There was a problem hiding this comment.
好的，已在e8297cb中提交了修改。
grassmann_tensor/tensor.py
Outdated
| U, S, Vh = torch.linalg.svd(tensor.tensor, full_matrices=full_matrices) | ||
|
|
||
| k = min(tensor.tensor.shape[0], tensor.tensor.shape[-1]) | ||
| k_index = tensor.tensor.shape.index(k) |
There was a problem hiding this comment.
分别进行svd后，需要允许有cut dimension的操作，这个在tn中很常见。大概就是删掉最小几个singular value，只保留最大的若干个，这个个数使用参数传进来，默认不进行cut，这里两个分块需要分别cut。
There was a problem hiding this comment.
好的，已在e8297cb中提交了修改。
- Perform SVD separately on even/odd parity blocks instead of entire
rank-2 tensor
- Update the test case with new implementation
Previously, the SVD was applied directly to the full Grassmann tensor,
which ignored the parity block structure and produced incorrect
decompositions. Now the tensor is split into even/odd blocks before
performing SVD, then recombined via block_diag to ensure correct
parity preservation.
Signed-off-by: Gausshj <[email protected]>
- Correct cutoff logic and support None as no-trunction mode - Add parameterized test cases for svd - Fix missing coverage on exception and boundary paths Signed-off-by: Gauss <[email protected]>
- Resolve coverage issues of svd function - Correct type annotations in test cases - Reformat the code of svd Signed-off-by: Gauss <[email protected]>
gausshj
force-pushed
the
dev/add-support-for-svd
branch
from
8bb8f29 to
e8297cb
Compare
hzhangxyz
requested changes
Oct 21, 2025
| self, | ||
| free_names_u: tuple[int, ...], | ||
| *, | ||
| full_matrices: bool = False, # When full_matrices=True, the gradient with respect to U and Vh will be ignored |
There was a problem hiding this comment.
full matrics可以直接是False，tensor下面，不会用到full matrics的svd。
|
|
||
| tensor = tensor.reshape((left_dim, right_dim)) | ||
|
|
||
| tensor.update_mask() |
| keep_even = torch.ones_like(S_even, dtype=torch.bool, device=S_even.device) | ||
| keep_odd = torch.ones_like(S_odd, dtype=torch.bool, device=S_odd.device) | ||
| else: | ||
| S_cat = torch.cat([S_even, S_odd]) |
There was a problem hiding this comment.
你的想法是对的，最普适的svd确实是应该这么做，不过我们需要broadcast的svd，所以如果直接设置一个cutoff会使得每个batch出现不同的S even/S odd选择方案。所以我们应该在S even里选择最大的若干个，S odd里选择最大的若干个。他们之间互不影响。这个函数的cutoff类型可以是None | int | tuple[int, int]
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.