Stability AI has recently released Stable Video Diffusion (SVD), an open-source text-to-video model that can generate short, high-quality video clips from images. The model can produce 14 or 25 frames at 576 x 1024 resolution, and it offers multi-view generation, 3D scene understanding, and camera control via LoRA. The SVD models are finetuned on 1M high-quality video samples, and the platform also introduced a Stable 3D model and Stable Doodle, a sketch-to-image tool. The release has garnered attention for its advanced capabilities and the ability to bring memes to life.
The Stable Video Diffusion demo on Hugging Face is 🤯 #stablediffusion #aivideo #urbanphotography https://t.co/RngPqq64R6
Memes come to life with stable video diffusion – mind-blowing! 🔥 https://t.co/slnehPQGgS
Stability AI has introduced new enhancements to its Stable Diffusion platform, most notably a new Stable 3D model. #StableDiffusion #DataScience #ArtificialIntelligence https://t.co/ny2mSq0UHQ
Stable Video Diffusion is out and it's ZOOMIN'! We upgraded our ComfyUI platform so you can try out the new AI video model made by the good folks at Stability Thanks to the incredible work by ComfyUI team it works fast on all sizes of our servers. Give it a shot! https://t.co/fRNwWbkOII
Stability AI releases Stable Doodle, a sketch-to-image tool https://t.co/2WEZWZNNxO Visit https://t.co/l8fNQzV9nN for more AI news. #AI #artificialintelligence #images #stabilityai
Stability just dropped Stable Diffusion Video less than 72 hours ago, and it's quite something else. Unbelievably advanced text-to-video/image-to-video capabilities, all from open-source. 10 stunning examples and how to run your own or access hosted options:
https://t.co/SawspKhC6R has just introduced Stable Video Diffusion, which allows you to do short animations from still images. This cat I made on DALL-E 3 can now move a bit https://t.co/AQIAxpa9aB https://t.co/xMntCYwUsa
Stable Video Diffusion (SVD): Scaling Latent Video Diffusion Models to Large Datasets In their paper, they mentioned 5 types of models they've created: High-Resolution Text-to-Video Model: Finetuned on 1M high-quality video samples, resolution 576×1024 for 50k iterations, batch… https://t.co/EDytRLk2lQ https://t.co/ZLu0vzGrKq
Available for research purposes, Stable Video Diffusion (SVD) models can produce short, high-quality clips from images. https://t.co/TWJrxwS5Kh
Stability just released Stable Diffusion Video this week. It's quite something. The text-to-video model is fully open source and let's you generate 14 or 25 frames at 576 x 1024. It can also do do multi-view generation, 3D scene understanding, and camera control via LoRA.… https://t.co/DiLByIAB3m https://t.co/Ic1KARd40q