Diffuser

Diffusion Model Overview

• Reference: Diffuers Quicktour - HuggingFace

• Reference：Diffusion Models for Machine Learning Introduction - AssemblyAI

Reference: How physics advanced generative AI - AssemblyAI

Diffusers by HuggingFace

Reference: HuggingFace - Understanding pipelines, models and schedulers.

Tasks Overview

• Unconditional image generation
• Text-to-image
• Image-to-image
• Inpainting
• Text / image -to-video
• Depth-to-image

Further Readings

Before original paper: 2020 Denoising Diffusion Probabilistic Models.

Mathematic explained is here: What are Diffusion Models?

Famous Diffusion Model

• OpenAI - DALL-E2. Explained of DALL-E2 from AssemblyAI: How DALL-E2 actually works.

  Now DALL-E3 is coming.

• Stable Diffusion by researchers and open-sourced by HuggingFace. Explained from HuggingFace is here: How does stable diffusion work.

  Overview of Stable Diffusion pipeline by HuggingFace is here.

  Origin paper of Text-to-image task: 2021 High-Resolution Image Synthesis with Latent Diffusion Models. (See page in HuggingFace)

/ML/