Latent Diffusion: The Latent Grid Size
Beginner
The problem, from first principles
Latent Diffusion (Rombach et al., 2022) runs the diffusion process in the compressed latent space of an autoencoder rather than on full-resolution pixels. A factor-8 autoencoder turns a 512 image into a 64 by 64 latent grid.
From theory to code
Implement latent_shape(h, w, f), the spatial size of the latent after downsampling by f.
Constraints
- Integer division.
Hints
Hint 1
Floor-divide each side by the factor.
Keywords
research-paperscomputer-visiondiffusionlatent
Python
Run your code to see results
No results yet
Run your code to check it against the examples, or submit to run every test.