A Blog post by IBM Granite on Hugging Face

HuggingFace's platform is a resource for developers and researchers working in natural language processing (NLP) and machine learning, offering insights into NLP models, tools, and datasets. Through articles, tutorials, and open-source projects, HuggingFace offers insights into state-of-the-art NLP techniques, transformer architectures, and transfer learning methods. Developers can learn about using pre-trained models, fine-tuning strategies, and deploying NLP applications with HuggingFace's libraries and APIs.

Hugging Face

IBM Research has released Mellea 0.4.0 and three Granite Libraries (granitelib-core-r1.0, granitelib-rag-r1.0, granitelib-guardian-r1.0) for building structured AI workflows on top of IBM Granite models. Mellea is an open-source Python library that replaces probabilistic prompt behavior with structured, maintainable generative programs using constrained decoding and composable pipelines. The 0.4.0 release adds native integration with the Granite Libraries, an instruct-validate-repair pattern via rejection sampling, and observability hooks. The Granite Libraries are collections of LoRA adapters fine-tuned for specific tasks like query rewriting, hallucination detection, and policy compliance, targeting the granite-4.0-micro model.

What's New in Mellea 0.4.0 + Granite Libraries Release

Overview Mellea 0.4.0 What Are the Granite Libraries Getting Started Mellea Granite Libraries Learn More Overview