Hi Community,
Here a result i am glad to share while testing the new kid in town of the Ai graphic tool.
This recent colab can be found at following url : https://colab.research.google.com/github/multimodalart/latent-diffusion-notebook/blob/main/Latent_Diffusion_LAION_400M_model_text_to_image.ipynb
Processing Note:
To produce this picture,, no init image were used.
- My prompt was : A Graphic Novel image by Moebius, colored by Marvel, of a close-up monster face of winter.
- Parameter were : size of 448x384, 60 steps, diversity 6, ddim (no PLMS).
Anyway you have to play around with those parameters to get result. The prompt is very sensitive for the results and it's better to get a 16GB GPU under the hood to run the script. Altough everyone speaks today about the release of the new Dall-e 2 from OpenAI which is stunning, but probably under restricted access, this new free colab is really interesting. I will see how far i can drive this one to a reliable comic style.
My previous posts are here.
Now also some periodical chirping news via https://twitter.com/Dbddv01
My Greetings to all the Alien Art Community for their encouragement.