The ability to seamlessly generate infinite remote sensing scenes

[02DF]

Dear Dr. Liu and team,
Thank you for your excellent work on Text2Earth! I'm particularly interested in the Unbounded Remote Sensing Scene Construction feature described in your paper README.md:67-68.
Would it be possible to share:
The trained Text2Earth model for unbounded scene generation, or Training code/methodology for the unbounded generation capability.
The ability to seamlessly generate infinite remote sensing scenes while maintaining visual coherence is groundbreaking for our research applications. Any guidance on implementation details or access to the model would be tremendously helpful.
Best regards.

[Chen-Yang-Liu]

Dear Dr. Liu and team, Thank you for your excellent work on Text2Earth! I'm particularly interested in the Unbounded Remote Sensing Scene Construction feature described in your paper README.md:67-68. Would it be possible to share: The trained Text2Earth model for unbounded scene generation, or Training code/methodology for the unbounded generation capability. The ability to seamlessly generate infinite remote sensing scenes while maintaining visual coherence is groundbreaking for our research applications. Any guidance on implementation details or access to the model would be tremendously helpful. Best regards.

You can use the Text2Earth-inpainting model to implement Unbounded Remote Sensing Scene Construction in an Outpainting way. This way is a common method to generate large images, for example: https://github.com/lkwq007/stablediffusion-infinity