v3.1.2

Features

• [FEAT] Any horizon explanation @marcopeix (#1393)

Bug fixes

• [FIX] Remove torch fix @elephaint (#1392)

v3.1.1

Hotfix

• [FIX] Backwards compatibility with saved models breaks if "explain" attribute isn't present @elephaint (#1389)

v3.1.0

New features

• [FEAT] (Shapley) explanations for univariate forecast models @elephaint @marcopeix (#1377)
• [FEAT] Any horizon prediction @elephaint @JQGoh (#1368)
• [FEAT]: Training window filtering @marcopeix (#1344)
• [FEAT] xLSTM @elephaint (#1363)

Bug fixes

• [FIX]: torch version @marcopeix (#1380)
• [FIX] Fixes for using uv @elephaint (#1364)
• [FIX] Bugs ISQF @elephaint (#1358)
• Update BiTCN doc dropout to match code @LemuelKL (#1376)
• Adds support for static variables in predict_insample method @nasaul (#1349)
• rm unused y_insample arg from pytorch losses @deven367 (#1360)

General

• Dev environment installation using uv @JQGoh (#1362)
• Enhancements @deven367 (#1371)
• Codespace setup for Neuralforecast @JQGoh (#1382)
• migrate tests from nbs to pytest @deven367 (#1353)

Documentation

• New documentation pipeline @deven367 (#1369)

v3.0.2

Enhancements

• Distributional predictions in predict_insample() @janovergoor (#1309)
• Optimizations in tsdataset: reduce allocations for large datasets @tylernisonoff (#1335)

Fixes

• [FIX]: Add logic to load custom models when using ReduceLROnPlateau @marcopeix (#1340)
• [FIX]: Fixes incorrect cuts in conformal prediction with conformal_error @elephaint (#1331)

v3.0.1

Features

• FEAT: Select basis functions in NBEATS @tblume1992 @marcopeix (#1191)
• FEAT: Add flash-attention @LeonEthan (#1295)
• FEAT: HuberIQLoss @elephaint (#1307)

Bug Fixes

• FIX: Fix iPython version @elephaint (#1282)
• FIX: Recurrent predictions @elephaint (#1285)
• FIX: Fix poor performance with the NegativeBinomial DistributionLoss @JQGoh (#1289)
• FIX: Add exclude_insample_y param to TimeXer for model loading @marcopeix (#1306)
• FIX: Set 2.0.0<=pytorch<=2.6.0 to avoid conflicts with networkx with Python 3.9 @marcopeix (#1318)
• FIX: Create windows once @elephaint (#1325)
• FIX: Add h_train to RNNs & fix issue with input_size @elephaint (#1326)
• FIX: Allow static vars only with NBEATSx and exogenous block @marcopeix (#1319)

v3.0.0

New features

• FEAT: TimeXer @marcopeix (#1267)
• All losses compatible with all types of models (e.g. univariate/multivariate, direct/recurrent) OR appropriate protection added.
• DistributionLoss now supports the use of quantiles in predict, allowing for easy quantile retrieval for all DistributionLosses.
• Mixture losses (GMM, PMM and NBMM) now support learned weights for weighted mixture distribution outputs.
• Mixture losses now support the use of quantiles in predict, allowing for easy quantile retrieval.
• Improved stability of ISQF by adding softplus protection around some parameters instead of using .abs.
• Unified API for any quantile or any confidence level during predict for both point and distribution losses.

Enhancements

• [DOCS] Docstrings @elephaint (#1279)
• FIX: Minor bug fix in TFT and a nicer error message for fitting with the wrong val_size @marcopeix (#1275)
• [FIX] Adds bfloat16 support @elephaint (#1265)
• Recurrent models can now produce forecasts recursively or directly.
• IQLoss now gives monotonic quantiles
• MASE loss now works

Breaking Changes

• [FIX] Unify API @elephaint (#1023)
• RMoK uses the revin_affine parameter instead of revine_affine. This was a typo in the previous version.
• All models now inherit the BaseModel class. This changes how we implement new models in neuralforecast.
• Recurrent models now require an input_size parameter.
• TCN and DRNN are now window models, not recurrent models
• We cannot load a recurrent model from a previous version to v3.0.0

Bug Fixes

• [FIX] Multivariate models give error when predicting when n_series > batch_size @elephaint (#1276)
• [FIX]: Insample predictions with series of varying lengths @marcopeix (#1246)

Documentation

• [DOCS] Update documentation @elephaint (#1274)
• [DOCS] Add example of modifying the default configure_optimizers() behavior (use of ReduceLROnPlateau scheduler) @JQGoh (#1015)

v2.0.1

Enhancements

• FEAT: Custom RNN layers for TFT @Yanam24 (#1230)
• FEAT: Add the horizon weighing to the distribution losses @mwamsojo (#1233)

Documentation

• DOCS: Add citation note @elephaint (#1244)
• fix: azul @AzulGarza (#1245)

v2.0.0

Breaking Change

• breaking: remove deprecated behavior @jmoralez (#1220)

v1.7.7

Bug Fixes

• [FIX] Backward compatibility: missing prediction_intervals @JQGoh (#1224)

v1.7.6

New Features

• [FEAT]: Support providing DataLoader arguments to optimize GPU usage @jasminerienecker (#1186)
• [FEAT]: Set activation function in GRN of TFT @marcopeix (#1175)
• [FEAT]: Conformal Predictions in NeuralForecast @JQGoh (#1171)

Bug Fixes

• [FIX]: Ability load models saved using versions before 1.7 @tylernisonoff (#1207)
• [FIX]: Conformal prediction issues @elephaint (#1179)
• [FIX]: Feature importance when using only hist_exog in TFT fails @elephaint (#1174)
• [FIX]: Remove unused output layer NBEATSx @elephaint (#1168)
• [FIX]: Fix Tweedie loss @elephaint (#1164)
• [FIX]: MLPMultivariate incorrect static_exog parsing @elephaint (#1170)
• [FIX]: Deprecate activation functions for GRU @marcopeix (#1198)

Documentation

• [DOC]: Tutorial on cross-validation @marcopeix (#1176)
• [DOC]: Build docs on release only @elephaint (#1183)