Thank you for sending your enquiry! One of our team members will contact you shortly.
Thank you for sending your booking! One of our team members will contact you shortly.
Duration 14 hours
Course Outline
Foundations of Multimodal LLMs in Vertex AI
- Exploring the multimodal capabilities within Vertex AI.
- Introducing Gemini models and their supported data modalities.
- Examining key use cases in enterprise and research environments.
Preparing the Development Environment
- Configuring Vertex AI settings for multimodal workflows.
- Managing datasets that span multiple data types.
- Practical Lab: Setting up the environment and preparing datasets.
Mastering Long Context Windows and Reasoning
- Understanding the mechanics of long-context workflows.
- Applying these capabilities to planning and decision-making scenarios.
- Practical Lab: Executing long-context analysis tasks.
Designing Cross-Modal Workflows
- Integrating analysis of text, audio, and image data.
- Linking multimodal steps to form cohesive pipelines.
- Practical Lab: Constructing a comprehensive multimodal pipeline.
Optimizing Gemini API Parameters
- Configuring efficient multimodal input and output handling.
- Enhancing inference speed and operational efficiency.
- Practical Lab: Fine-tuning Gemini API configurations.
Advanced Applications and System Integrations
- Building interactive multimodal agents and assistants.
- Connecting with external APIs and third-party tools.
- Practical Lab: Developing a complete multimodal application.
Evaluation Strategies and Iterative Improvement
- Testing the performance of multimodal systems.
- Defining metrics for accuracy, alignment, and model drift.
- Practical Lab: Assessing and refining multimodal workflows.
Conclusion and Future Directions
Requirements
- Strong command of Python programming.
- Practical experience in developing machine learning models.
- Working knowledge of handling multimodal data types, including text, audio, and images.
Target Audience
- AI Researchers
- Senior/Advanced Developers
- Machine Learning Scientists
