Possible Enhancements/optimizations. - Is quantization possible for kohya? ai toolkit has it. Serious speed boost for 1024 training, and bucketing multi resolutions.

1. Is it possible to add quantization? with ai toolkit i get 22.1gb vram doing 1024 training, and 1400 steps done in 45 mins
in kohya 1024 training, takes me nearly 2 hours with the same amount.

2. Multiresolution training with bucketing, it buckets all my images into 512, 768, and 1024.

overall I'm getting quite good results.

The issue names no files, tests, or entry points. Start by reviewing the existing training and resolution-bucketing flow, then compare it with the ai toolkit's quantization and bucketing behavior. Done means the requested quantization and multiresolution bucketing are supported with improved training speed or VRAM use.

Written by the indexing model from the issue text.