From ecbbcd5cc5b9484f4de910a5d7b673589cff8038 Mon Sep 17 00:00:00 2001 From: Fangzhou Ai <31551580+Fangzhou-Ai@users.noreply.github.com> Date: Sat, 3 Oct 2026 05:01:54 +0000 Subject: [PATCH] [DeepSeek][ROCm] V4.1-Flash: recommend the ROCm 10.0 nightly image Point the AMD install image at vllm/vllm-openai-rocm:nightly-rocm100 instead of :nightly. The ROCm 10.0 nightly is the image the MI355X DeepSeek-V4.1-Flash InferenceX AgentX lane runs and validates on. Signed-off-by: Fangzhou Ai <31551580+Fangzhou-Ai@users.noreply.github.com> Co-authored-by: Cursor --- models/deepseek-ai/DeepSeek-V4.1-Flash.yaml | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/models/deepseek-ai/DeepSeek-V4.1-Flash.yaml b/models/deepseek-ai/DeepSeek-V4.1-Flash.yaml index c07a4dae6..ec64f9d13 100644 --- a/models/deepseek-ai/DeepSeek-V4.1-Flash.yaml +++ b/models/deepseek-ai/DeepSeek-V4.1-Flash.yaml @@ -5,7 +5,7 @@ meta: provider: "DeepSeek" description: "DeepSeek V4.1 Flash vision-language MoE (552B backbone; 8B active per prompt token, 16B per output token) combining sliding-window plus compressed sparse attention with a two-level indexer, engram n-gram memory, hyper-connections, and a DSpark multi-token draft head." date_added: 2026-09-09 - date_updated: 2026-10-02 + date_updated: 2026-10-03 difficulty: advanced tasks: - text @@ -37,8 +37,9 @@ model: nightly_required: true default_frontend: rust docker_image: - # Includes the AITER mHC delayed-pre path from vllm-project/vllm#56503. - amd: "vllm/vllm-openai-rocm:nightly" + # ROCm 10.0 nightly channel; includes the AITER mHC delayed-pre path from + # vllm-project/vllm#56503. + amd: "vllm/vllm-openai-rocm:nightly-rocm100" architecture: moe parameter_count: "552B" active_parameters: "8-16B" @@ -537,7 +538,7 @@ guide: | landed on `main` in [vllm-project/vllm#56228](https://github.com/vllm-project/vllm/pull/56228), so any nightly from 2026-09-10 on serves it, and the launch-day `deepseekv41-flash-0909` tag is no longer needed. On AMD it is - `vllm/vllm-openai-rocm:nightly`, which also carries + `vllm/vllm-openai-rocm:nightly-rocm100`, the ROCm 10.0 nightly, which also carries [vllm-project/vllm#56503](https://github.com/vllm-project/vllm/pull/56503) — the change that moves the mHC delayed pre block off the eager Torch reference and onto AITER and postdates the 0909 tag.