Update Docker Docs and Examples (Published Images, GGUF, Qwen3 TTS)

[xashr]

Summary

This PR now updates the Docker documentation and examples to reflect the current state:

• Reference published images from ghcr.io/0xshug0/audio.cpp as primary source for images, local builds still supported
• Switch to GGUF models from https://huggingface.co/audio-cpp/audio.cpp-gguf
• Add Qwen3 TTS as second example to demonstrate voice cloning
• Add ToC, clean up, etc.

[xashr]

@0xShug0 Currently we have /Docker.md and /docs/docker.md . We should only maintain/keep one of them. Any preference? Keep Docker.md or switch to /docs/docker.md ?

Unfortunately I also noticed that the Qwen3 TTS3 does not work with CUDA in the container. Not related to this PR, but might be a regression from #84 . It's due to torch_random.cpp and #ifdef GGML_USE_CUDA . ==> Created #109

ggml_cuda_init: found 1 CUDA devices (Total VRAM: 32107 MiB):
  Device 0: NVIDIA GeForce RTX 5090, compute capability 12.0, VMM: yes, VRAM: 32107 MiB
load_backend: loaded CUDA backend from /app/libggml-cuda.so
load_backend: loaded CPU backend from /app/libggml-cpu-alderlake.so
audiocpp_cli failed: Qwen3 TTS generation requires a build with CUDA support

[0xShug0]

@xashr Let's use docs/docker.md

I also noticed the regression because of #102. I merged the PR last night. Did you try the latest commit?

[xashr]

I see you merged PR #109 already. Great, that fixed the Qwen3 TTS example.

README.md references docs/docker.md now.

[0xShug0]

@xashr Merged! Thanks!