Best practices for scanning and processing

HI! Very lovely work, many thanks for your contribution to the field 🙏

Couple of questions
- Do you have some recommendations on expected frame rate for video processing? Are we talking 1/2hz?
- Im looking to reduce the GPU memory need for large datasets (e.g 4000) - how would you tackle this? I've tried batching depth estimation using known camera poses and put the GLB together at the end, but no dice

350 frames all in one (like intended)

350 frames batched in 50 slices

Thank you!

This is a discussion about frame rates and GPU memory for large video scans, but it names no file, test, or entry point. Start by reviewing the repository's documented scanning and processing guidance. Done would require a clear, documented recommendation for expected frame rates and handling large datasets.

Written by the indexing model from the issue text.