README: add a community Android (LiteRT GPU) port to Deployment

[john-rocky]

Thanks for the Deployment block in the Tools section. This adds a fourth step to it for an Android build of D-FINE-S: the dfine-small-coco weights converted to LiteRT (.tflite) and running fully on the phone GPU through the LiteRT CompiledModel API, with a Kotlin sample app.

• Pixel 8a: both graphs fully GPU-resident (511/511 and 850/850 ops); about 450 ms per frame end-to-end, so still-image use rather than real-time. The number is in the line so nobody expects otherwise.
• Detections match PyTorch at IoU 0.99–1.00 on a COCO val image, with matching class and score.
• Apache-2.0 is kept, and the zoo section links back to this repo.

Zoo section: https://github.com/john-rocky/LiteRT-Models#d-fine-s — weights: https://huggingface.co/litert-community/D-FINE-S-LiteRT

Happy to reword or drop it if the block should stay ONNX/TensorRT only. Thanks again.