models

Why we included this project

ONNX exists so a model trained in one framework can run in another without retraining, and this zoo shows what that looks like in practice. It gathers pre-trained vision, NLP, and generative models, covering everything from image classification and object detection to translation, into one browsable collection. Because each entry is a well-known architecture exported to the portable ONNX standard, the repo doubles as a working reference for judging how easily a model will move across runtimes. One honest caveat: the project is now preserved for historical purposes, with new model sharing moved to the Hugging Face onnxmodelzoo hub, so treat these artifacts as a useful starting point rather than a live registry.