Completed
13:46 New workflows: Qwen Image, OmniGen 2, Cosmo, HighDream
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
ComfyUI Models 2025 - The Best Model for Photos and Art - SDXL vs Flux vs SD3.5
Automatically move to the next video in the Classroom when playback concludes
- 1 0:00 Introduction and overview of all models
- 2 0:58 Stable Diffusion 1.5 basics and workflow
- 3 2:30 Noise levels, resizing, and VRAM usage
- 4 4:31 Stable Diffusion 2.1 improvements and differences
- 5 5:52 SDXL leap: dual encoders and 1024 resolution
- 6 7:33 SDXL base vs refiner workflow explained
- 7 8:03 Running SDXL and results demonstration
- 8 8:57 Stable Diffusion 3 and 3.5: MMDiT and rectified flow
- 9 9:45 SD 3.5 text encoders and typography support
- 10 10:46 Flux models: Def, Chel, and Chroma variants
- 11 12:26 Flux context: powerful in-place editing with text
- 12 13:46 New workflows: Qwen Image, OmniGen 2, Cosmo, HighDream
- 13 14:55 Qwen Image for text and layout generation
- 14 16:00 OmniGen 2: assistant-like editing and composition
- 15 17:25 Cosmo Predict 2: Nvidia model for physical accuracy
- 16 18:34 HighDream models: mixture of experts approach
- 17 20:51 PixArt Sigma: direct 4K generation and installation steps
- 18 25:03 HunyuanDiT: multilingual support and simplicity
- 19 26:07 Stable Cascade: three-stage generation approach
- 20 27:51 Playground v2.5: SDXL-style alternative
- 21 28:48 Conclusion and final thoughts
