Open-sourcing Knowledge Distillation Code and Weights of SD-Small and SD-Tiny

OpenAI has released the code and weights for its knowledge distillation models, SD-Small and SD-Tiny, aimed at providing lightweight alternatives to larger models. SD-Small has 2.7 billion parameters, while SD-Tiny contains 1.1 billion parameters, both optimized for efficiency in deployment. This release enables practitioners to leverage smaller models with reduced computational requirements, facilitating broader accessibility and application of AI in resource-constrained environments.