Document or link to existing best practices

To make our job easier, we should probably make a page that lists useful external resources on best practices for storing foundation model weights and embeddings. E.g.:

Foundation Models:
- Store as .pt2 or .safetensors?
- https://github.com/developmentseed/pixelverse/issues/32
- https://wherobots.com/solution-gallery/rasterflow-bring-your-own-model/#2-pytorch-2-export-formats

Embeddings:
- https://geoembeddings.org/bestpractices.html

Visualization:
- https://github.com/developmentseed/pixelverse/issues/27

Anything else in the end-to-end pipeline to document? I can start a skeleton page with some of these as a start, and we can gradually fill them in. Key thing is to not duplicate or overlap with the scope of https://geoembeddings.org, so maybe focus on more technical/implementation details/decisions we're making.

Start by reviewing the linked Pixelverse issues, PyTorch export-format resource, and geoembeddings best-practices page, then check the project’s existing documentation structure. Create a page covering implementation-focused best practices for model weights, embeddings, visualization, and other pipeline decisions without duplicating geoembeddings.org; the listed resources should be linked and organized clearly.

Written by the indexing model from the issue text.