Add aspect ratio bucketing to training scripts

**Is your feature request related to a problem? Please describe.**
When fine tuning SDXL, images are required to be a fixed size (1024x1024) which involves a lot of cropping that both takes time/resources, and often causes important parts of the image to get cropped out, which lowers model quality.

**Describe the solution you'd like.**
The ideal solution would be a simple option for user to enable aspect ratio bucketing (e.g. a command argument `--enable-bucketing`) that will let them train with multiple image sizes

Start by locating the SDXL training scripts and the code responsible for enforcing fixed image sizes and cropping. Define how the optional --enable-bucketing argument should select multiple image sizes, then verify that training can preserve varied aspect ratios without requiring every image to be cropped to 1024x1024.

Rédigé par le modèle d'indexation à partir du texte de l'issue.