I have lots of thoughts on the current state of txt2vid, and most aren’t kind. But Kilng’s ability to take direction from source images is downright impressive.
Prompt: “Three characters argue. The left tries to calm the situation while the right holds back the middle character”
January 21, 2025 at 1:33 PM
I have lots of thoughts on the current state of txt2vid, and most aren’t kind. But Kilng’s ability to take direction from source images is downright impressive.
Prompt: “Three characters argue. The left tries to calm the situation while the right holds back the middle character”
Same prompt, two different stable diffusion models. My custom model on the left and protogeninfinity on the right. The abstraction of faces and clothing come from the model, making it easy to direct everything but the aesthetics. Those stay consistently abstract and amazing thanks to the model.
July 2, 2023 at 6:26 AM
Same prompt, two different stable diffusion models. My custom model on the left and protogeninfinity on the right. The abstraction of faces and clothing come from the model, making it easy to direct everything but the aesthetics. Those stay consistently abstract and amazing thanks to the model.