Update noisy latent examples.

noisy_latent_composition/README.md

@@ -5,11 +5,10 @@ You can Load these images in [ComfyUI](https://github.com/comfyanonymous/ComfyUI
 Here are examples of Noisy Latent Composition. Noisy latent composition is when latents are composited together while still noisy before the image is fully denoised. Since general shapes like poses and subjects are denoised in the first sampling steps this lets us for example position subjects with specific poses anywhere on the image while keeping a great amount of consistency.
 
 Here is an example. This example contains 4 images composited together. 1 background image and 3 subjects.
-The total steps is 16. The latents are sampled for 4 steps with a different prompt for each. The background is 1280x704 and the subjects are 256x512 each. After these 4 steps the images are still extremely noisy. The subjects are then composited (pasted) onto the background with some feathering applied. The rest of the sampling steps are then run on this composited image.
+The total steps is 16. The latents are sampled for 4 steps with a different prompt for each. The background is 1920x1088 and the subjects are 384x768 each. After these 4 steps the images are still extremely noisy. The subjects are then composited (pasted) onto the background with some feathering applied. The rest of the sampling steps are then run on this composited image.
 
-I also do a second pass to increase the resolution from 1280x704 to 1920x1088 but that is optional.
 
-These examples are done with the AnythingV3 model.
+These examples are done with the WD1.5 beta 3 illusion model.
 
 ![Example](noisy_latents_3_subjects.png)