darts · Issues· 219 open
Open on GitHubLocally synced open issues (discussions stay on GitHub)
- #2968
multi_models = True and output_chunk_length > 1 : index future covariates relatively to the forecasted step
feature requestcore improvementUpdated Sep 18, 2026 - #3200
[New Model] TiRex-2 zero-shot multivariate forecasting
new modelUpdated Sep 18, 2026 - #3179
Support cold-start forecasting for Torch models without target history
Updated Aug 19, 2026 - #3177
Safe-by-default model loading
core improvementUpdated Aug 19, 2026 - #3128
[QUESTION] Have you considered more advanced outlier detection and value imputation?
feature requestcore improvementUpdated Aug 17, 2026 - #2842
Add One-Hot Encoder to Time Axis Encoders
feature requestpr_welcomeUpdated Aug 14, 2026 - #3171
Allow TFTExplainer to explain more series than the model batch size
Updated Aug 7, 2026 - #2685
Add support for TFT feature importance over time
feature requestpr_welcomeUpdated Aug 1, 2026 - #2489
TFTExplainer: getting explanations when trained on multiple time series
buggood first issueUpdated Jul 30, 2026 - #2758
[FEAT] RIN for covariates
feature requestpr_welcomeUpdated Jul 29, 2026 - #1649
Retrieve last hidden state for RNNModel and BlockRNNModel
good first issueimprovementUpdated Jul 22, 2026 - #2043
future_covariate support to NBEATS & NHits models
feature requestUpdated Jul 15, 2026 - #3140
[Feature] Efficient DTW and clustering
triageUpdated Jun 20, 2026 - #2463
TFT Model resume training from checkpoint not matching continuous training
Updated Jun 8, 2026 - #2359
[New Model] Pre-trained, "foundational" models
new modelpr_welcomeUpdated Jun 4, 2026