-
Notifications
You must be signed in to change notification settings - Fork 17
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
release 0.10.2 #1008
Merged
Merged
release 0.10.2 #1008
Conversation
This file contains 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
Draft fuinction that worked on my initial testing using OpSo 0.9.1, but still needs testing on 0.10.1
BCEWithLogitsLoss_hot and CrossEntropyLoss_hot now accept **kwargs in __init__ and pass to the init of the parent class
the BasePreprocessor class now implements saving the time it takes to do each preprocessing step in the sample's `.runtime` attribute, a dictionary indexed like the preprocessor's .pipeline needs tests
Feat_profile_preprocessing
Feat set seed
resolves #961 BasePreprocessor._generate_sample, if passed pd.Series, should use the class method so that the resulting sample contains the labels rather than just path/start/end time
in bioacoustics_model_zoo.load()
resolves matplotlib not being released when saving spectrogram to file #987
…953 avoid modifying original overlay_df object passed to Overlay.__init__()
scale DC and Nyquist frequency by just /N not 2/N
cast np int/float types to native types for timedelta also fixed in two other places, one in audio.py and one in localization.py
resolve spectrum estimation 'wrong' #947
resolves issue 924 futurewarning pandas
BoxedAnnotations.from_raven_files raven_files and audio_files arguments can now be a path (str or pathlib.Path) rather than list of paths.
for convencience, predict(samples) wraps str or path into a 1-item list
resolves one_hot_clip_labels breaks after using convert_labels on BoxedAnnotations object #916 BoxedAnnotations had a _spawn method but was calling its init method instead with just the df argument at the end of convert_labels. This caused loss of other attributes. Now calls _spawn to retain all attributes not explicitly passed. Added check within test_convert_labels.
resolves random_trim_audio in default SpectrogramPreprocessor doesn't work #904
doesn't implement **kwargs option, but resolves the immediate bad-practice code of specifying dataloader, wandb_session, and progress_bar as positional arguments rather than keyword arguments. This will avoid silent failuer if user implements different __call__ method with different arguments
resolves avoid mismatch between model.device='string' and torch.device #888 allows us to convert string to torch.device when user changes the value eg m.device='cpu'
this was added accidentally at some point
automatically sets `torch.multiprocessing.set_sharing_strategy("file_system")` during opensoundscape import. We may want to revisit this decision, but it seems that this is the recommended setting for avoiding issues seen when using parallelized DataLoader see discussion and recommended solution here pytorch/pytorch#11201 (comment)
improve AudioTrim action
adds @sanruizguz and @LeonardoViottio to author list!
made some small edits to notes on downloading files
device handling has changed, but now ScoreCAM is fixed and we can use any device instead of just boolean use_cuda flag
resolves some issues with efficientnet hashes. Also, update cnn_architecture tests to assert type. Fix bug where architecture was not returned after initializing.
avoids ImportError: cannot import name 'ErrorString' from 'docutils.core' when using docutils 0.20.1
louisfh
approved these changes
Jun 3, 2024
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good - made a few small comments.
adds some extra documentation of functions and tests renames a test in test_utils from test_cnn to test_set_seed_results_in_deterministic_weights_cnn_init adds tests for utils.cast_np_to_native adds Most Recent Batch Loss as well as logging Epoch Running Average Loss (formerly called DistLoss) during training step logging of training progress
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
This release includes bug fixes and minor improvements to 0.10.1. This branch off of develop was created for the 0.10.2 release and does not include breaking changes that will be incorporated in the 0.11.0 release.