Noising and denoising process, Forward vs. reverse diffusion, Mathematical foundations

Latent diffusion architecture, Image generation pipeline, Key components: VAE, U-Net, CLIP

Generating images from text prompts, Prompt engineering techniques, Negative prompting and guidance scale

Transformer architecture refresher, Few-shot learning capabilities, Fine-tuning vs. prompting

Building custom chat applications, RLHF (Reinforcement Learning from Human Feedback), Safety and alignment techniques

Codex & CodeGen architectures, Prompt patterns for code generation, Evaluation metrics (pass@k)

DALL-E 2/3 architecture, Stable Diffusion XL features, Commercial vs. open-source options

Synthesizing video clips from text, Temporal attention mechanisms, Current limitations and challenges

In-context learning mechanisms, Prompt engineering for few-shot, Zero-shot capabilities analysis