[Bug]: [Bug][Docker] Recipe references vllm/vllm-openai-rocm:deepseekv41-flash-0909 but the image is not available on Docker Hub

### Your current environment

AMD MI308X

### 🐛 Describe the bug

First, a huge thank you to the entire vLLM engineering team and community. Your work on making high-performance LLM serving accessible across NVIDIA, AMD ROCm, Intel, TPU, and CPU backends is genuinely transformative — the pace of model coverage (DeepSeek V4, V4.1-Flash, Kimi-K3, GLM-5.3, …), the unified frontends (Python + Rust), the disaggregation and KV-cache innovations all land in months, not years. The DeepSeek-V4.1-Flash support that merged today in PR #56201 is fantastic — thank you for sustaining this project as a true open-source effort and supporting the AMD ROCm path so thoroughly.

I'm reaching out to flag a reproducible inconsistency between the official recipe documentation and what is actually published on Docker Hub, which is currently blocking a production rollout on our side.

Problem

The official recipe for DeepSeek-V4.1-Flash on AMD MI300X ( https://recipes.vllm.ai/deepseek-ai/DeepSeek-V4.1-Flash?hardware=mi300x ) instructs users to pull:

`docker pull vllm/vllm-openai-rocm:deepseekv41-flash-0909`

However, this tag does not exist on Docker Hub. Attempts to pull it from multiple mirrors consistently fail with:

Error response from daemon: pull access denied … repository does not exist or may require 'docker login' (404 / manifest unknown)
Error response from daemon: received unexpected HTTP status: 403 Forbidden
The same 403 / not-found responses are returned when retrying through Chinese registry mirrors (docker.1ms.run, docker.m.daocloud.io, dockerproxy.net, docker.xuanyuan.me).
This blocks following the recipe exactly as written, and is misleading for users who arrive at the recipe page expecting an official one-line deployment path.

where can I find the `vllm/vllm-openai-rocm:deepseekv41-flash-0909` ？

thank you again~~~

3× AMD Instinct MI308X nodes (gfx942, ROCm), 192 GiB HBM each
OS: Ubuntu 22.04, kernel 6.8, Docker 24.x
ROCm driver: 6.2.4