foundation-models 5
- Meta Llama 3.1 Architecture: 405B Scaling, 4D Parallelism, and DPO Alignment
- The 2024 Foundation Model Wave: Architectural Shifts in Training, Data, and Serving
- Chameleon Architecture: Mixed-Modal Early-Fusion and Discrete Token Generation
- Generative AI Taxonomy: Core Mechanisms from Transformers to Flow Matching
- Deconstructing LLaMA 3 Architecture: GQA, SwiGLU, and Tokenizer Scaling