Phase 12

Multimodal AI

Models that see, hear, read, and reason across modalities.

01Vision Transformers and the Patch-Token PrimitiveLearn~120 minutes 02CLIP and Contrastive Vision-Language PretrainingBuild~180 minutes 03From CLIP to BLIP-2 — Q-Former as Modality BridgeBuild~180 minutes 04Flamingo and Gated Cross-Attention for Few-Shot VLMsLearn~120 minutes 05LLaVA and Visual Instruction TuningBuild~180 minutes 06Any-Resolution Vision: Patch-n'-Pack and NaFlexBuild~120 minutes 07Open-Weight VLM Recipes: What Actually MattersLearn + lab~180 minutes 08LLaVA-OneVision: Single-Image, Multi-Image, Video in One ModelBuild~180 minutes 09Qwen-VL Family and Dynamic-FPS VideoLearn~120 minutes 10InternVL3: Native Multimodal PretrainingLearn~120 minutes 11Chameleon and Early-Fusion Token-Only Multimodal ModelsBuild~180 minutes 12Emu3: Next-Token Prediction for Image and Video GenerationLearn~120 minutes 13Transfusion: Autoregressive Text + Diffusion Image in One TransformerBuild~180 minutes 14Show-o and Discrete-Diffusion Unified ModelsLearn~120 minutes 15Janus-Pro: Decoupled Encoders for Unified Multimodal ModelsBuild~120 minutes 16MIO and Any-to-Any Streaming Multimodal ModelsLearn~120 minutes 17Video-Language Models: Temporal Tokens and GroundingBuild~180 minutes 18Long-Video Understanding at Million-Token ContextBuild~180 minutes 19Audio-Language Models: the Whisper to Audio Flamingo 3 ArcBuild~180 minutes 20Omni Models: Qwen2.5-Omni and the Thinker-Talker SplitBuild~180 minutes 21Embodied VLAs: RT-2, OpenVLA, π0, GR00TLearn~180 minutes 22Document and Diagram UnderstandingBuild~180 minutes 23ColPali and Vision-Native Document RAGBuild~180 minutes 24Multimodal RAG and Cross-Modal RetrievalBuild~180 minutes 25Multimodal Agents and Computer-Use (Capstone)Capstone~240 minutes

“Stay hungry, stay foolish.”

— Steve Jobs