braindecode 1.6.1

What's Changed

• [MNT] Bump version to 1.6.0dev0 for next development cycle by @bruAristimunha in #1024
• Fix batch=1 BatchNorm error in ContraWR, DeepSleepNet, and EEGMiner by @bruAristimunha with @Copilot in #1028
• Add optional revision support to pull_from_hub and forward it to HF snapshot download by @PierreGtch with @Copilot in #1029
• Add monthly automated PyPI stable release workflow by @bruAristimunha in #1030
• Make monthly release tag-only with a version-bump PR by @bruAristimunha in #1031
• Fix monthly-release workflow parse error (empty expression in comment) by @bruAristimunha in #1032
• Add InterpolatedEEGPT model by @PierreGtch in #1038
• docs: fix incorrect import path in CONTRIBUTING.md (model.base -> models.base) by @YihengLi-1 in #1033
• docs: fix EEGClassifier typo (EEGClassifer -> EEGClassifier) in cropped decoding example by @YihengLi-1 in #1034
• docs: add Yneuro logo to landing page institution strip by @bruAristimunha in #1041
• ci: cap pytest below 9.1 to keep pytest_cases working by @adammounir in #1047
• docs: fix broken EEGNeX quickstart snippet on the landing page by @bkowshik in #1048
• Remove deprecated aliases overdue since v1.4 (EEGNetv4, SleepStagerEldele2021, TSceptionV1, BNCI2014001) by @bkowshik in #1045
• Track network-dependent tests with pytest markers by @bruAristimunha with @Copilot in #1052
• Add EEG-DINO foundation model by @bruAristimunha in #1049
• fix: apply safe standalone MNE preprocessors via callable (#885) by @YihengLi-1 in #1037
• ci: install CPU-only torch in CI to fix disk-exhaustion crash by @bkowshik in #1054
• Warn when drop_last silently discards all training batches on small datasets by @bkowshik in #1053
• [MAINT] Silence drop_last small-trainset warning in signal-arg tests by @adammounir in #1056
• [FIX] Align EEGTCNet, SSTDPN & ATCNet with original source implementations by @bruAristimunha in #1064
• Fix SSL tutorial PCA plot to render distinct sleep-stage colors by @PierreGtch with @Copilot in #1063
• Add TCFormer (Temporal Convolutional Transformer) model by @bruAristimunha in #1065
• clarify model summary type labels by @tayal-sarthak in #1044
• Enhance README generation with dataset license inference by @Fashad-Ahmed in #1014
• Add STEEGFormer model (#1040) by @adammounir in #1042
• Model & filter fidelity fixes (#1067–#1071) + AugmentedDataLoader expansion & augmentation speedups by @bruAristimunha in #1073
• feat: opt-in channel positions in the batch for heterogeneous montages by @bruAristimunha in #1072
• fix: wrap aux-return standalone MNE preprocessors (closes #1055) by @bruAristimunha with @Copilot in #1074
• BrainModule: add the (missing) optional spatial Fourier ChannelMerger + fix float-dilation crash by @bruAristimunha in #1076
• refactor(modules): share sinusoidal PE primitive + add GatedLinearUnit (GEGLU) by @bruAristimunha in #1078
• feat(docs): pin braindecode installation version in generated notebooks by @Fashad-Ahmed in #1080
• Add a Modality column to the model summary table by @bkowshik in #1079
• Expose FBMSNet pre-classifier features by @bruAristimunha with @Copilot in #1083
• Add causal filtering support to FilterBankLayer by @bruAristimunha with @Copilot in #1084
• Fix SincShallowNet MPS backward failure by @bruAristimunha with @Copilot in #1087
• Fix EEGSym reference architecture by @bruAristimunha in #1089
• feat: add MVPFormer foundation model by @bruAristimunha in #1057

New Contributors

• @YihengLi-1 made their first contribution in #1033
• @adammounir made their first contribution in #1047
• @bkowshik made their first contribution in #1048

Full Changelog: v1.5.1...v1.6.1