[Parity with vLLM, SGLang, ATOM]: Public Nightly NGC docker images too

[functionstackx]

🚀 The feature, motivation and pitch

AMD ATOM, vLLM (CUDA & AMD), SGLang (CUDA & AMD) all public nightly docker images. Can TRTLLM have public nightly ngc docker image builds too?

https://hub.docker.com/r/rocm/atom-dev

https://hub.docker.com/r/vllm/vllm-openai/tags
https://hub.docker.com/r/vllm/vllm-openai-rocm/tags
https://hub.docker.com/r/lmsysorg/sglang-rocm/tags
https://hub.docker.com/r/lmsysorg/sglang/tags?page=2

+viz @xinli-sw @ncomly-nvidia @kedarpotdar-nv

Alternatives

No response

Additional context

end user wastes couple hours building image by themselves

Before submitting a new issue...

• Make sure you already searched for relevant issues, and checked the documentation and examples for answers to frequently asked questions.