First of all, I did manage to get through all the issues I've been experiencing previously and everything is working great.
I can confirm no VRAM issues, both for the small model and the large model (I'm using the gradio app and I'm on a 4090).
The only problem I'm experiencing is that the image to video--while it generates some great videos--doesn't seem to reflect the prompt at all. For example I have an image of a boy standing, and I entered the prompt "The boy is walking". I tried many many times but never got one video that actually had the boy walking.
Which makes me wonder, am I doing something wrong? Is image-to-video prompting working for everyone else?
First of all, I did manage to get through all the issues I've been experiencing previously and everything is working great.
I can confirm no VRAM issues, both for the small model and the large model (I'm using the gradio app and I'm on a 4090).
The only problem I'm experiencing is that the image to video--while it generates some great videos--doesn't seem to reflect the prompt at all. For example I have an image of a boy standing, and I entered the prompt "The boy is walking". I tried many many times but never got one video that actually had the boy walking.
Which makes me wonder, am I doing something wrong? Is image-to-video prompting working for everyone else?