Skip to content

Commit

Permalink
Add CHANGELOG
Browse files Browse the repository at this point in the history
  • Loading branch information
ulupo committed May 14, 2024
1 parent 29b7642 commit 8f9a3c9
Show file tree
Hide file tree
Showing 4 changed files with 27 additions and 7 deletions.
28 changes: 24 additions & 4 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
@@ -1,13 +1,33 @@
## 0.1.0
# Release notes

<!-- do not remove -->

## 0.1.1
### Breaking Changes

- Store hard and soft losses as Python scalars instead of 0-dimensional NumPy arrays ([#3](https://github.com/Bitbol-Lab/DiffPaSS/issues/3))

### New Features

- Add `remove_groups_not_in_both` function ([876d017](https://github.com/Bitbol-Lab/DiffPaSS/commit/876d01792a0206ee209478bd2ee5a4c122f2ab9d))
- Unify type annotations for `group_sizes` ([#7](https://github.com/Bitbol-Lab/DiffPaSS/issues/7))

- Add possibility to include diagonals in `IntraGroupSimilarityLoss` computations ([#5](https://github.com/Bitbol-Lab/DiffPaSS/issues/5))

- Store hard and soft losses as Python scalars instead of 0-dimensional NumPy arrays ([#3](https://github.com/Bitbol-Lab/DiffPaSS/issues/3))

### Backward-incompatible changes
### Bugs Squashed

- Fix `fit_bootstrap` appending empty lists ([#1](https://github.com/Bitbol-Lab/DiffPaSS/issues/1))

>## 0.1.0
### Breaking Changes

- Change signature of `get_robust_pairs` ([ebb160c](https://github.com/Bitbol-Lab/DiffPaSS/commit/ebb160c512e6aed2cdb9865bdb9b2088a8e0ffd4))

## 0.0.1
### New Features

- Add `remove_groups_not_in_both` function ([876d017](https://github.com/Bitbol-Lab/DiffPaSS/commit/876d01792a0206ee209478bd2ee5a4c122f2ab9d))

>## 0.0.1
- First DiffPaSS public release
2 changes: 1 addition & 1 deletion diffpass/model.py
Original file line number Diff line number Diff line change
Expand Up @@ -593,7 +593,7 @@ class IntraGroupSimilarityLoss(Module):
relationships.
Similarity matrices are expected to be square and symmetric. Their diagonal
elements are ignored if `exclude_diagonal` is set to True.
elements are ignored if `exclude_diagonal` is set to ``True``.
If `group_sizes` is provided, the loss is computed by comparing the flattened
and concatenated upper triangular blocks containing intra-group similarities.
Otherwise, the loss is computed by comparing the upper triangular part of the
Expand Down
2 changes: 1 addition & 1 deletion nbs/model.ipynb
Original file line number Diff line number Diff line change
Expand Up @@ -1093,7 +1093,7 @@
" relationships.\n",
"\n",
" Similarity matrices are expected to be square and symmetric. Their diagonal\n",
" elements are ignored if `exclude_diagonal` is set to True.\n",
" elements are ignored if `exclude_diagonal` is set to ``True``.\n",
" If `group_sizes` is provided, the loss is computed by comparing the flattened\n",
" and concatenated upper triangular blocks containing intra-group similarities.\n",
" Otherwise, the loss is computed by comparing the upper triangular part of the\n",
Expand Down
2 changes: 1 addition & 1 deletion nbs/train.ipynb
Original file line number Diff line number Diff line change
Expand Up @@ -236,7 +236,7 @@
" # Check that the hard loss of the optimized permutation is close to the ground truth\n",
" assert np.abs(results.hard_losses[-2][-1] - hard_loss_identity_perm) < 1e-4\n",
"\n",
"test_information_bootstrap() "
"test_information_bootstrap()"
]
},
{
Expand Down

0 comments on commit 8f9a3c9

Please sign in to comment.