[AMD] Add MiniMax-M3-FP8 MI355X ATOMESH update 0623

[seungrokj]

Summary

• Eliminate all hardcoded MODEL_NAME == "DeepSeek-V4-Pro" / per-model checks from server_atom.sh
• All model-specific configuration (env vars, parallel flags, MTP flags, KV cache flags, HF overrides) now driven from models_atom.yaml using the same python3 yaml.safe_load pattern as server_vllm.sh
• Add MiniMax-M3-MXFP4 and MiniMax-M3-MXFP8 entries to models_atom.yaml with EAGLE3 MTP flags
• Image bump for minimaxm3-fp8-mi355x-atom-disagg: rocm/atom-dev:MiniMax-M3-20260622 → rocm/atom-dev:MiniMax-M3-20260623

Fields added to models_atom.yaml

Field	Purpose
env	Space-separated KEY=VALUE pairs exported unconditionally
tp_dp_flags	Parallel flags for TP+DPA mode
tp_dp_env	Env vars exported only in TP+DPA mode
ep_dp_flags	Parallel flags for EP+DPA mode
ep_dp_env	Env vars exported only in EP+DPA mode
mtp_flags	Flags prepended to SPEC_ARGS before $DECODE_MTP_SIZE
kv_cache_flags	Full --kv_cache_dtype flag string
hf_overrides	JSON string passed to --hf-overrides

PR Review Checklist

• Verified that as of the moment of typing this, this is the latest version of PR_REVIEW_CHECKLIST.md
• Verified that the general code quality meets the InferenceX standard and does not make the code quality any worse.
• Verified that this PR has passed PR validation. Please link to GitHub Action workflow that shows this.
• Verified that this PR passes evals. Please link to GitHub Action workflow that shows this.
• Verified that speculative decoding PRs uses chat templates to align the AL distribution to real world
• If a company claims that they support vLLM/SGLang as first class LLM inference engines on their hardware, I have verified that the respective vLLM/SGLang submission has been made before additional frameworks (TRT-LLM, ATOM, etc.). The only exceptions are for new hardware, such as MI455X UALoE72, Vera Rubin NVL72, Rubin NVL8, etc., and for new model architectures where there is an actual reason why vLLM/SGLang does not fundamentally support them yet.
• Verified that the single-node recipes are similar to the official vLLM recipes and/or the SGLang cookbook:
  • If they are not, I have verified that a PR has been opened in vLLM recipe repo or SGLang repo and linked it below in the additional detail section:
• If any of the above criteria cannot reasonably be satisfied, I have provided additional reasoning below.

🤖 Generated with Claude Code

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=28149790315
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=28149790315

[functionstackx]

@Oseltamivir can u review this? tho it seems like evals r failing potentially failing

[seungrokj]

@functionstackx @Oseltamivir let me first check something and will ping when it is ready!