Build Multimodal Models

Training programme overview

Practical exploration of multimodal models for vision, audio, video and image generation. Focuses on integrating multimodal data flows, compatible architectures, open-source libraries and concrete steps to assemble models and pipelines.

Survey of multimodal model types and concrete application scenarios.

Processing and fusion of vision, audio and video signals to feed unified models.

Techniques for conditional image generation and output post-processing.

Pipeline architecture: preprocessing, multimodal encoding, fusion and decoding.

Examples of practical tools and components for experimentation and prototyping.

Training · Advanced