- Sink In
- Posts
- Sora Competitor Kling AI Launches Globally
Sora Competitor Kling AI Launches Globally
Wassup - this is SinkIn Newsletter, a 5 minutes read made at sinkin.ai to cover the most interesting stuff in the Image AI world.
Kling AI, the Sora competitor from Chinese tech company Kuaishou, just launched its international version and people around the globe have been testing out its video generation capabilities. Results posted across X are mind-blowing. They are smooth, natural and simply stunning. The site offers some free credits upon signup which allows generating 6 videos. It takes around 2-3 minutes to generate one video now.
Videos generated by Kling AI
Tencent just released PhotoMaker V2 with improved ID fidelity. While it maintains the generation quality, editability, and compatibility with any plugins that PhotoMaker V1 offers, it provides scripts for integration with ControlNet, T2I-Adapter, and IP-Adapter to offer better control capabilities. Users can further customize scripts for upgrades, such as combining with LCM for acceleration or integrating with IP-Adapter-FaceID or InstantID to further improve ID fidelity. Try out the demo here.
Latest virtual try-on advancements use latent diffusion models for realistic garment inpainting, enhancing online shopping. However, existing methods lack flexible control over garments, faces, poses, and scenes. IMAGDressing-v1 addresses this with a garment UNet capturing semantic and texture features and a hybrid attention module for scene control. It supports plugins like ControlNet and includes the IGPair dataset with 300,000 clothing pairs. Extensive tests show state-of-the-art performance in human image synthesis. Try out the demo here.
At SinkIn we recently launched Drawbert, a product that aims to make it real easy to create portrait art. We recently added two new modes: converting pet photo to painting and converting portraits to Pixar style characters. With a few clicks, you can create timeless art for your furry friends or entertain your family with Pixar magic. Try them out and let us know what you think.
Stability AI just launched Stable Video 4D, an AI model that transforms a single video into multiple novel-view videos of eight different angles, in about 40 seconds. It was developed based on Stable Video Diffusion and Stable Video 3D. This new model enhances virtual try-ons, game development, and video editing by providing versatile, realistic perspectives. It's available on Hugging Face and aims to support future applications in VR and more.
Meme of the Week
That’s it for today, hope they were as good as my afternoon snack bar!
What'd you think of today's edition? |