Scalecast

Latest version: v0.19.10

Safety actively analyzes 723158 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 23 of 28

0.5.6

Added
- Added documentation on Read the Docs
- Added cilvel information to export functions
Changed
Fixed

0.5.5

Added
- Added CILevel info to export model_summaries function
Changed
Fixed
- Fixed an issue where plots were diplaying incorrect confidence levels if `cilevel` had been changed since training it

0.5.4

Added
- Added the rnn estimator
Changed
- plot_loss argument no longer considered a hyperparameter value for LSTM and RNN models
Fixed
- Fixed an issue where "==" wasn't being accepted in the `evaluated_as` argument in the `pop_using_criterion()` function
- Scaler in history saved as 'minmax' instead of None for LSTM and RNN models

0.5.3

Added
- EvaluatedModels to `__repr__()`
Changed
- Does not call `infer_freq()` as often, making code more efficient
Fixed
- sometimes the attribute `ci_bootstrap_samples` was being called `bootstrap_samples`, changed everything to `bootstrap_samples` only

0.5.2

Added
Changed
Fixed
- Fixed an error that occured when calling the `__repr__()` method if no models had been evaluated first

0.5.1

Added
- Added `export_fitted_vals()` function
- Added ci option to the `results_vis()` function in notebook
- Added the `get_funcs()` function
Changed
- No `Xvars` in LSTM model, changed to lags (now model will only look at its own history)
- No `normalizer` in LSTM model (always uses a minmax scaler now)
- LSTM model can no longer be tuned
- Got rid of all lstm model grids
- changed `__str__()` and `__repr__()` so that they now offer better info
Fixed
- Fixed the LSTM model by scaling the dependent variable and unscaling it (minmax) when it comes out and getting rid of other Xvars

Page 23 of 28

© 2025 Safety CLI Cybersecurity Inc. All Rights Reserved.