Discover published sets by community

Core features include gradient boosting algorithms, regularized boosting techniques, and tree pruning. Typical applications are structured data prediction challenges, such as Kaggle competitions, and classification problems requiring high performance.

Core features include gradient boosting framework using tree based learning algorithms, optimized for distributed and efficient training. Typical applications are similar to XGBoost, including large-scale machine learning.

Core features include tight integration with NumPy, transparent use of GPU, and efficient symbolic differentiation. Typical applications are deep learning models, research prototypes, and mathematical computations in Python.

Core features include scalable machine learning library, data parallelism and fault tolerance on commodity hardware. Typical applications are large-scale machine learning, supporting classification, regression, clustering, and topic modeling.

Core features include speed, modularity, and expressive architecture. Typical applications are convolutional networks for image classification, and multimedia projects.

Core features include dataflow graph computation, GPU/CPU computing, automatic differentiation. Typical applications are image recognition, natural language processing, and time series analysis.

Core features include a wide array of machine learning algorithms, pipelines, and data pre-processing methods. Typical applications are classification, regression, clustering, and dimensionality reduction.

Core features include dynamic computational graph ('define-by-run'), native support for GPU computation, and modularity. Typical applications are rapid prototyping, research and development, and deep learning applications.

Core features include simplifying training of neural networks, practical deep learning for coders, and a focus on modern best practices. Typical applications are educational resources and deep learning prototypes.