Playing around with RunwayML to breathe new life into some old SD 1.5 images. It’s nice, but $$$ to use. Playing around with some local toys to use on RunPod on the go forward. For now, though, enjoy!
Here is Amy on a video chat. It does well with the mouth movement and facial expressions I think:
It actually seems to do better with ‘realistic’ images vs these pseudo realistic ones, in terms of translating movement. This seems a bit stiff, but I do like the OG image set this is from, so I’ll probably play some more here:
The original image called for her hair to be up in a messy bun, or maybe a ponytail. In this video we attempt to have the robots let her hair down. Not great, but not the worst I’ve seen so far:
In this one she magically pulls a bottle out of her hair, because for all its wonders, AI remains kind of stupid:
‘Till next time, friends. Stay well!
Leave a Reply