Transformers, new features, transfer learning#424
Merged
Conversation
…scaler_unit_variance
…ault true and default false lists
Ipc should not be changed to AvgIpc like this because it would break all rdkit_raw models.
…th RobustScaler and PowerTransformer. Updated documentation in related sections. Added functions to ModelFileReader to read out transformer specific parameters. Changed models that test RobustScaler and PowerTransformer to use RF to speed up the training
…default function for all sklearn parameters
… it more generalizeable. Fixed tests. Fixed bug where the imputer_strategy parameter was not used
…e log scale features
…ndicator' flag because that changed the number of features and crashed.
…rming using SklearnPipelineWrapper
…g/AMPL into feat_scaled_rdkit_mordred
…e to transformers_x
…model, if transformers are saved and loaded correctly, and if transform_dataset_key_config is saved correctly
…r want to set that manually. Instead added a check when saving metadata to see if the parameters object has that attribute
…well as infill nan or extremely large values
Codecov Report❌ Patch coverage is
@@ Coverage Diff @@
## 1.8.0 #424 +/- ##
==========================================
+ Coverage 49.69% 51.61% +1.91%
==========================================
Files 37 38 +1
Lines 11717 11982 +265
==========================================
+ Hits 5823 6184 +361
+ Misses 5894 5798 -96
Flags with carried forward coverage won't be shown. Click here to find out more.
... and 2 files with indirect coverage changes 🚀 New features to boost your workflow:
|
…r molecular weight
… colliding split_uuids
…re_transformers when transformers is None. This does not test the pipeline with no transformers, just that the function returns correctly
- Tests that the heavyatom_col paramter is used correctly and cases when there is no heavyatom_col. - Tests that the NotImplementedError is raised correctly when there is no feature count or if there is no way to featurize data. - Tests that the Identity features transforms are returned correctly. And that an error is raised if an unrecognized feature transform is used.
…k and multitask models are trained using the same dataset, the scaled_descriptors copy of the featurized file will only contain response_columns for the single task models, and not all columns for the multitask models. This does not cause an issue during training, but when making predictions, the '_actual' columns won't exist. This causes the function to crash. This patch looks in the original dataset_key csv and finds the response columns and merges them into the scaled_descriptors file.
…t. Without this step, PowerTransformer failes
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
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.
This is a large pull request with 3 new features.