v1.7.0

[jmoralez]

New Features

• [FEAT] Added TSMixerx model @elephaint ([FEAT] Added TSMixerx model #921)
• Add Time-LLM @marcopeix (Add Time-LLM #908)
• [FEAT] Added TSMixer model @elephaint ([FEAT] Added TSMixer model #914)
• Add option to support user defined optimizer for NeuralForecast Models @JQGoh (Add option to support user defined optimizer for NeuralForecast Models #901)
• [FEAT] Added NLinear model @ggattoni ([FEAT] Added NLinear model #900)
• [FEAT] Added DLinear model @cchallu ([FEAT] Added DLinear model #875)
• support refit in cross_validation @jmoralez (support refit in cross_validation #842)
• use environment variable to get id as column in outputs @jmoralez (use environment variable to get id as column in outputs #841)
• support different column names for ids, times and targets @jmoralez (support different column names for ids, times and targets #838)
• polars support @jmoralez (polars support #829)
• add callbacks to auto models @jmoralez (add callbacks to auto models #795)

Bug Fixes

• [FIX] Avoid raised error for varied step_size parameter during predict_insample() @JQGoh ([FIX] Avoid raised error for varied step_size parameter during predict_insample() #933)
• [FIX] 926 auto ensure all models support alias and 924 Configuring hyperparameter space for Auto* Models @elephaint ([FIX] 926 auto ensure all models support alias and 924 Configuring hyperparameter space for Auto* Models #927)
• fix base_multivariate window generation @jmoralez (fix base_multivariate window generation #907)
• Fix optuna multigpu @jmoralez (Fix optuna multigpu #889)
• support saving and loading models with alias @jmoralez (support saving and loading models with alias #867)
• [FIX] Polars .columns produces list rather than Pandas Index @akmalsoliev ([FIX] Polars .columns produces list rather than Pandas Index #862)
• add missing models to filename dict @jmoralez (add missing models to filename dict #856)
• ensure exogenous features are lists @jmoralez (ensure exogenous features are lists #851)
• fix save with save_dataset=False @jmoralez (fix save with save_dataset=False #850)
• copy config in optuna @jmoralez (copy config in optuna #844)
• Fixed: Exception: max_epochs is deprecated, use max_steps instead. @twobitunicorn (Fixed: Exception: max_epochs is deprecated, use max_steps instead. #835)
• fix single column 2d array polars df @jmoralez (fix single column 2d array polars df #830)
• move scalers to core @jmoralez (move scalers to core #813)
• [FIX] Default AutoPatchTST config @cchallu ([FIX] Default AutoPatchTST config #811)
• [FIX] ReVin Numerical Stability @dluuo ([FIX] ReVin Numerical Stability #781)
• On Windows, prevent long trial directory names @tg2k (On Windows, prevent long trial directory names #735)

Documentation

• removed documentation for missing argument @yarnabrina (removed documentation for missing argument #913)
• feat: Added cross-validation tutorial @MMenchero (feat: Added cross-validation tutorial #897)
• chore: update license to apache-2 @AzulGarza (chore: update license to apache-2 #882)
• [FEAT] Model table in README @cchallu ([FEAT] Model table in README #880)
• redirect to mintlify docs @jmoralez (redirect to mintlify docs #816)
• add missing models to documentation @jmoralez (add missing models to documentation #775)

Dependencies

• add windows to CI @jmoralez (add windows to CI #814)
• address future warnings @jmoralez (address future warnings #898)
• use scalers from coreforecast @jmoralez (use scalers from coreforecast #873)
• add python 3.11 to CI @jmoralez (add python 3.11 to CI #839)

Enhancement

• Reduce device transfers @elephaint (Reduce device transfers #923)
• extract common methods to BaseModel @jmoralez (extract common methods to BaseModel #915)
• remove TQDMProgressBar callback @jmoralez (remove TQDMProgressBar callback #899)
• use fsspec in save and load methods @jmoralez (use fsspec in save and load methods #895)
• Feature/Check input for NaNs when available_mask = 1 @JQGoh (Feature/Check input for NaNs when available_mask = 1 #894)
• switch flake8 to ruff @Borda (switch flake8 to ruff #871)
• use future instead of deprecation warnings @jmoralez (use future instead of deprecation warnings #849)
• add frequency validation and futr_df debugging methods @jmoralez (add frequency validation and futr_df debugging methods #833)

---

This discussion was created from the release v1.7.0.