Add Qwen3.5 with llama.cpp validated path to text-to-text Android LP

[BelachewRachel]

Adds the validated Qwen3.5 llama.cpp path to the text-to-text Android Learning Path.

Updates:

• Adds Qwen3.5 Q4_K_M llama.cpp to the validated model table.
• Adds llama.cpp / GGUF guidance for the custom-model path.
• Adds instructions to build the official llama.cpp Android AAR, apply the Qwen validated adapter example, download the GGUF model package, and run the model on Android.
• Updates the overview wording so the LP covers ExecuTorch, LiteRT, ONNX Runtime GenAI, and llama.cpp validated paths.

---

Before submitting a pull request for a new Learning Path, please review Create a Learning Path

• I have reviewed Create a Learning Path

Please do not include any confidential information in your contribution. This includes confidential microarchitecture details and unannounced product information.

• I have checked my contribution for confidential information

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of the Creative Commons Attribution 4.0 International License.

[matt-cossins]

Discussed on slack