Support JPEG and PNG image inputs

Add dedicated text and image embedding endpoints

Refactor hub storage to use Hugging Face dataset repository

added hub storage and patch embeddings

added tiles

Expose GeoText embeddings and region boxes

Refactor type hints in handler.py and infer.py for improved consistency

Fix RGB data handling in fetch_rgb_geotiff function

Add root endpoint for health check

Add six library to requirements.txt for dependency management

Refactor type hints in server.py for improved clarity

Enhance INFERENCE.md with detailed build and usage instructions

Refactor Docker setup for GeoText-1652 inference endpoint

Add GeoText inference endpoint support

Duplicate from truemanv5666/GeoText1652_model