LLMOps with OpenVINO: Converting Florence-2 Models to IR

Deploying modern vision-language models efficiently requires understanding model optimization and lightweight inference runtimes. This text-based course guides you through core LLMOps principles and the OpenVINO toolkit, focusing on converting and running Florence-2 models for fast, efficient execution. You will build a firm grasp of model graph optimization, runtime execution, and precision reduction techniques through clear written explanations and code snippets. What you'll learn: - Understand fundamental LLMOps concepts and runtime optimization principles - Convert Florence-2 vision-language models to OpenVINO Intermediate Representation (IR) format - Apply model quantization and precision tuning to minimize memory footprint - Configure inference pipelines for multimodal vision and language tasks - Benchmark inference latency and resource efficiency across hardware - Implement automated workflows for model export and verification. The curriculum starts with essential definitions and architecture overviews before advancing into step-by-step conversion workflows and practical optimization exercises. Designed for beginners in AI deployment, this course requires no prior experience with OpenVINO or LLMOps pipelines. Start reading today to master model optimization and deploy high-performance vision-language models.