Media Synthesis Museum restores early genAI models like Big Sleep
multimodalart · x · 2026-08-27
The Media Synthesis Museum on Hugging Face aims to preserve generative cultural objects. They have restored early classic models like Big Sleep, VQGAN+CLIP, and CLIP Guided Diffusion to run on modern hardware or ZeroGPU, eliminating dependency on legacy Colab notebooks.
More from Multimodal
- Pika Labs API demonstrates pasta cooking video generation — aziz4ai · 2026-08-27
- MiniMax H3 Generates Nike Air Max Commercial with Prompt — aziz4ai · 2026-08-27
- Seamless H3 Video Join with Tone Compensation Node — rkfg_me · 2026-08-27
- Seedance 2.5 demo turns an entire desert into a sleeping goddess — umesh_ai · 2026-08-27
- Creating Kpop songs and reactive MVs with MiniMax Music 3 — SIR_NVAX_A_LOT · 2026-08-27
- User creates seamless animation using H3 Max and Seedream 5 Pro in 5 mins — techhalla · 2026-08-27