Hugging Face

The Hugging Face Hub is the central platform for accessing and sharing AI assets such as models, datasets, and Spaces, and for using an ecosystem of community and developer contributions.

Transformers is Hugging Face’s library for building and fine-tuning state-of-the-art models for NLP and related transformer-based workloads.

Datasets provides tools to access and process datasets for machine learning training and evaluation.

Spaces are app-style deployments/demos hosted on the Hugging Face Hub, enabling teams to share interactive experiences alongside model and dataset artifacts.

Inference Providers gives developers access to model inference through a unified API surface for multiple model providers.

Inference Endpoints provides managed, production deployment for selected models, enabling organizations to serve models in their own infrastructure context.

Text Generation Inference is an optimized inference toolkit used to serve text generation workloads efficiently.

AutoTrain helps automate parts of model training workflows for teams who want to train models with reduced engineering effort.

Infinity is described by Hugging Face as a product designed to increase the speed of an in-production model’s data processing pipeline.