11 repos
Python libraries and tools for building, converting, and standardizing deep learning models across TensorFlow, PyTorch, JAX, and other frameworks. This cluster centers on reducing friction between incompatible ML ecosystems—enabling tensor operations, model interchange, dataset handling, and runtime compilation that work across multiple backends. Central projects like Keras (a high-level API), Ivy (a unified framework abstraction layer), and IREE (compiler infrastructure) exemplify the cluster's focus on making deep learning code portable and framework-agnostic.