24 repos
Libraries and utilities for publishing, managing, and loading machine learning models using PyTorch's model hub mixin interface and the SafeTensors format. This cluster focuses on standardized model serialization, distribution, and compatibility across PyTorch ecosystems, with particular emphasis on speech and audio models (dialect and language identification via Whisper and MMS architectures). Repositories here provide tooling to make models discoverable, reproducible, and interoperable across different deployment contexts.