Decision 3.0: open, multimodal SOTA decision models from 0.6B to 27B (huggingface.co)

Hugging Face has unveiled Decision 3.0, a significant update in their series of open multimodal foundation decision models, ranging from 600 million to 27 billion parameters. This advancement introduces models optimized for zero-shot classification tasks, demonstrating enhanced performance and versatility within the AI/ML community. The release includes several versions, such as the 26B and 8B models, catering to various infrastructure capabilities and application needs. The Decision 3.0 models are noteworthy for their ability to process and integrate multiple data modalities, expanding the potential use cases in areas such as natural language processing and computer vision. With a strong emphasis on accessibility, these models are expected to accelerate innovation in AI-driven applications, allowing developers to leverage state-of-the-art capabilities without extensive resources. This release represents a significant step towards democratizing advanced decision-making models, making sophisticated AI technologies more widely available for both research and practical applications.