Torchfitter

Fixed

- Solve doc typos.
- Fix logger and trainer tests.
- Fix incomplete `quickstart` in docs.
- Fix logging bug in `GPUStats` callback.

Added

- Add support for mixed precision training.
- Add ElasticNet regularization.
- Add testing methods and their tests: `check_monotonically_decreasing` and `compute_forward_gradient`.
- Add cuda seed setting in Manager.
- Add option to only use deterministic algorithm in the Manager class.

Changed

- Update logo and README.
- Update tests with new testing methods.
- Make some method on Trainer and Manager private.

Fixed

- Solve bug in callbacks where the handler was not calling in appropiate order.

Removed
- Remove `ElasticNet` regularization because the implementation was not correct.

Changed

- Change `params_dict` in the Trainer to a specific class that tracks the internal state.
- Change README.
- Update tests.
- Change logic of TQDM to be updated in each batch instead of in each epoch.
- Change optimization loop to be of type `condition-loop` instead of `iteration-loop`. This is, the loop is now a `while` loop.

Added

- Add `Manager` class to handle multiple experiments.
- Add support for computing metrics in the optimization loop via `torchmetrics`.
- Add `GPUStats`, `ReduceLROnPlateau` and `ProgressBarLogger` callbacks.
- Add testing utility to check gradients: `compute_forward_gradient`.
- Add more functions to `utils`: `FastTensorDataLoader`, `check_model_on_cuda`.

Fixed

- Solve warning where learning rate scheduler was being called before loss.

Changed

- Change `_compute_penalty` in favour of `compute_penalty`.
- Change `_train` in favour of `train_step`.
- Change `_validate` in favour of `validation_step`.
- Update tests to be correct.

Added

- Added new `reset_parameters` method in the trainer.
- Added requirements file for example.
- Added `trainer` example in `.py` format.
- Added `manager.ipynb` example.

Fixed

- Fix error in setup naming.
- Fix moving the tensors to device. Now, it is done in each batch.

Changed

- Change the `requirements.txt` to remove unnecessary dependencies.

Added

- Added possibility to use L1 and ElasticNet [regularization](https://github.com/Xylambda/torchfitter/pull/3).
- Added new testing module.
- Added tests for the new functionalities.

Changed

- Updated README to add brief tutorial on how to create regularization algos.
- Updated tests for trainer.

Fixed

- Fixed minor typos in README.

Added

- Added badges from [shield.io](https://shields.io/)