Skip to content

Feature/ Find largest batch size for TorchForecastingModel #2905

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
daidahao wants to merge 9 commits into
base: master
Choose a base branch
from
Open

Feature/ Find largest batch size for TorchForecastingModel #2905

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
86f5e67
Add `scale_batch_size()` to `TorchForecastingModel`
daidahao Sep 18, 2025
4f0bb57
Add a `scale_batch_size()` test and update dataloader-related tests
daidahao Sep 18, 2025
c1fc1b9
Update CHANGELOG & `scale_batch_size` docstring
daidahao Sep 18, 2025
0f72406
Fix a bug in `_setup_for_train` when `val_dataset` is `None`
daidahao Sep 19, 2025
9a73510
Add tests to validate unchanged model weights after training
daidahao Sep 19, 2025
c8823d7
Merge branch 'master' into feature/torch-scale-batch-size
daidahao Sep 19, 2025
4241a01
Merge branch 'master' into feature/torch-scale-batch-size
daidahao Sep 21, 2025
e2381f2
Merge branch 'master' into feature/torch-scale-batch-size
daidahao Sep 29, 2025
7ede1c9
Fix on `load_best` when merging commits
daidahao Sep 29, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,7 @@ but cannot always guarantee backwards compatibility. Changes that may **break co

**Improved**

- Added `scale_batch_size()` method to torch-based models to find the largest batch size. It is recommended to re-initialize the model after calling this method before calling `fit()`. [#2905](https://github.com/unit8co/darts/pull/2905) by [Zhihao Dai](https://github.com/daidahao).
- Added hyperparameter `skip_interpolation` to `TFTModel` that will replace 1D interpolation on feature embeddings with linear projection. When `True`, it can greatly increase training and inference efficiency while predictive accuracy remains largely unaffected. [#2898](https://github.com/unit8co/darts/pull/2898) by [Zhihao Dai](https://github.com/daidahao).
- Added mixed precision and 16-bit precision support to `TorchForecastingModel`. Simply specify `{"precision": "bf16-mixed" }` for `pl_trainer_kwargs` to enable mixed precision training. Alternatively, declare a custom `pytorch_lightning.Trainer` with a `"precision"` parameter and pass the trainer to `fit()`. Other precision options such as `"64-true"` and `"16-mixed"` supported by `pytorch_lightning` are also allowed. [#2883](https://github.com/unit8co/darts/pull/2883) by [Zhihao Dai](https://github.com/daidahao).
- 🔴 Added future and static covariates support to `BlockRNNModel`. This improvement required changes to the underlying model architecture which means that saved model instances from older Darts versions cannot be loaded any longer. [#2845](https://github.com/unit8co/darts/pull/2845) by [Gabriel Margaria](https://github.com/Jaco-Pastorius).
Expand Down
Loading