A slightly overcooked test lora for cosmos 7B, trained on single images.
Get the text to video example from: https://comfyanonymous.github.io/ComfyUI_examples/cosmos/
For best results set: width and height to 1024 and length to 1. Use "LoraLoaderModelOnly" to load the lora.
Use the positive prompt: "cute anime girl with massive fennec ears and a big fluffy tail long blonde wavy hair blue eyes"