Machine Learning Libraries
scikit-learn, gradient boosting, deep-learning frameworks and model formats.
This is an overview — pick a sub-topic below.
Sub-topics (13)
scikit-learn
The standard library for classic ML: preprocessing, models, pipelines and evaluation with one consistent API.
XGBoost, LightGBM and CatBoost
Gradient-boosted decision trees — top performers on tabular data.
PyTorch
The most popular deep-learning framework: dynamic graphs, Pythonic API, huge ecosystem.
TensorFlow and Keras
Google's framework (TensorFlow) and its high-level API (Keras 3, which also runs on JAX and PyTorch).
JAX, Flax, Haiku and Equinox
JAX's composable transformations with neural-network libraries built on top.
MXNet, PaddlePaddle and MindSpore
Other deep-learning frameworks: Apache MXNet (now retired), Baidu's PaddlePaddle, Huawei's MindSpore.
ONNX and ONNX Runtime
A standard model format and a fast cross-platform inference engine.
Optuna, Ray Tune, Hyperopt and scikit-optimize
Hyperparameter optimisation libraries using smart search strategies.
AutoGluon, FLAML, H2O AutoML, TPOT and PyCaret
AutoML libraries that build strong models with a few lines of code.
SHAP, LIME, InterpretML, Captum and Alibi
Explainability toolkits for tabular, image and text models.
River
Online machine learning — models that learn one sample at a time from streams.
sktime, Darts and Nixtla
Time-series forecasting, classification and anomaly detection libraries.
PyMC, NumPyro, Pyro, Stan, TensorFlow Probability and emcee
Probabilistic programming: describe a model with distributions and infer parameters with MCMC or variational inference.
Spotted a mistake or want to add detail? Suggest an improvement — Content Contributors' edits are reviewed by editors before they go live.
Discussion (0)
Sign in or join free to ask questions and discuss this topic.