bloc97

Yeah, I'll get back to working on it during the week. Adding a more intuitive prompt editing method, such as using brackets (eg. `a landscape of (mountains:-10) with (rivers and...

I believe the easiest way for someone to get started is to simply download the diffusers library and run this jupyter notebook. While it does require the diffusers library, it...

@sunwoo76 Awesome! I didn't look very deeply in the code but it looks good at first glance from reading the README. I think showing a visualization of the attention maps...

You can try the method in `InverseCrossAttention_Release.ipynb`. However, this works very well only with images that were generated with stable diffusion. For real images, inversion with high CFG is an...

Examples where inversion doesn't work well: Left is original image, right is reconstruction with prompt. https://www.pexels.com/photo/tray-of-pumpkins-on-a-knitted-sweater-5429788/ ```python gen_latents = inversestablediffusion(input_image, "pumpkins in a tray seen from above", refine_iterations=10, refine_skip=0.4, guidance_scale=3.0)...

> Have you thought about whether fine tuning the model or using an embedding (like textual inversion or dreambooth) That might actually work! A better reconstruction usually allows for better...

@hmartiro There's Google's Imagic paper that just got released. https://arxiv.org/abs/2210.09276 From the paper, it seems that inverting the prompt embeddings too (not just the latent) yields even better results. In...

I'm not sure we would want a !pip install inside of a jupyter notebook. It is fine to use in colab as the environment is already set up, but using...

I will need to see whether this can be optimized for 6GB, as we need to save the attention maps, which is pretty big for the lower parts of the...

While cross attention control does help in prompt editing, it is not perfect, some changes will be present due to how diffusion models work. What is the output for a...