Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/34j/lightgbm-callbacks
A collection of LightGBM callbacks. (DART early stopping, tqdm progress bar)
https://github.com/34j/lightgbm-callbacks
dart early-stopping hacktoberfest lgbm lightgbm lightgbm-dart scikit-learn sklearn sklearn-compatible tqdm
Last synced: 2 months ago
JSON representation
A collection of LightGBM callbacks. (DART early stopping, tqdm progress bar)
- Host: GitHub
- URL: https://github.com/34j/lightgbm-callbacks
- Owner: 34j
- License: mit
- Created: 2023-05-05T02:58:45.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-10-21T17:08:59.000Z (2 months ago)
- Last Synced: 2024-10-22T06:41:33.649Z (2 months ago)
- Topics: dart, early-stopping, hacktoberfest, lgbm, lightgbm, lightgbm-dart, scikit-learn, sklearn, sklearn-compatible, tqdm
- Language: Python
- Homepage:
- Size: 408 KB
- Stars: 3
- Watchers: 3
- Forks: 0
- Open Issues: 12
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- Contributing: CONTRIBUTING.md
- Funding: .github/FUNDING.yml
- License: LICENSE
Awesome Lists containing this project
README
# LightGBM Callbacks
A collection of [LightGBM](https://github.com/microsoft/LightGBM) [callbacks](https://lightgbm.readthedocs.io/en/latest/Python-API.html#callbacks).
Provides implementations of `ProgressBarCallback` ([#5867](https://github.com/microsoft/LightGBM/pull/5867)) and `DartEarlyStoppingCallback` ([#4805](https://github.com/microsoft/LightGBM/issues/4805)), as well as an `LGBMDartEarlyStoppingEstimator` that automatically passes these callbacks. ([#3313](https://github.com/microsoft/LightGBM/issues/3313), [#5808](https://github.com/microsoft/LightGBM/pull/5808))
## Installation
Install this via pip (or your favourite package manager):
```shell
pip install lightgbm-callbacks
```
## Usage
### SciKit-Learn API, simple
```python
from lightgbm import LGBMRegressor
from lightgbm_callbacks import LGBMDartEarlyStoppingEstimator
from sklearn.datasets import load_diabetes
from sklearn.model_selection import train_test_split
X, y = load_diabetes(return_X_y=True)
X_train, X_test, y_train, y_test = train_test_split(X, y)
LGBMDartEarlyStoppingEstimator(
LGBMRegressor(boosting_type="dart"), # or "gbdt", ...
stopping_rounds=10, # or n_iter_no_change=10
test_size=0.2, # or validation_fraction=0.2
shuffle=False,
tqdm_cls="rich", # "auto", "autonotebook", ...
).fit(X_train, y_train)
```
### Scikit-Learn API, manually passing callbacks
```python
from lightgbm import LGBMRegressor
from lightgbm_callbacks import ProgressBarCallback, DartEarlyStoppingCallback
from sklearn.datasets import load_diabetes
from sklearn.model_selection import train_test_split
X, y = load_diabetes(return_X_y=True)
X_train, X_test, y_train, y_test = train_test_split(X, y)
X_train, X_val, y_train, y_val = train_test_split(X_train, y_train)
early_stopping_callback = DartEarlyStoppingCallback(stopping_rounds=10)
LGBMRegressor(
).fit(
X_train,
y_train,
eval_set=[(X_train, y_train), (X_val, y_val)],
callbacks=[
early_stopping_callback,
ProgressBarCallback(early_stopping_callback=early_stopping_callback),
],
)
```
### Details on `DartEarlyStoppingCallback`
Below is a description of the `DartEarlyStoppingCallback` `method` parameter and `lgb.plot_metric` for each `lgb.LGBMRegressor(boosting_type="dart", n_estimators=1000)` trained with entire `sklearn_datasets.load_diabetes()` dataset.
| Method | Description | iteration | Image | Actual iteration |
| ---------- | -------------------------------------------------------------------------------------------- | ----------------------------------------------------------- | ------------------------------------- | ---------------- |
| (Baseline) | If Early stopping is not used. | `n_estimators` |  | 1000 |
| `"none"` | Do nothing and return the original estimator. | `min(best_iteration + early_stopping_rounds, n_estimators)` |  | 50 |
| `"save"` | Save the best model by deepcopying the estimator and return the best model (using `pickle`). | `min(best_iteration + 1, n_estimators)` |  | 21 |
| `"refit"` | Refit the estimator with the best iteration and return the refitted estimator. | `min(best_iteration, n_estimators)` |  | 20 |
## Contributors ✨
Thanks goes to these wonderful people ([emoji key](https://allcontributors.org/docs/en/emoji-key)):
This project follows the [all-contributors](https://github.com/all-contributors/all-contributors) specification. Contributions of any kind welcome!
