---
cited_by:
  - '[[thoughts/inference/deepseek-v4.1-serving#111-vl|dsv41 §1.11]]'
  - '[[thoughts/papers/deepseekai2026deepseekv41flashpushinglimitskv#2-v4--v41-diff|dsv41 §2]]'
  - '[[thoughts/inference/deepseek-v4.1-serving#5-sharp-edges-bugs-and-fixes-that-expose-invariants|dsv41 §5]]'
date: '2026-11-11'
kind: pr
local:
  - file:///Users/aarnphm/workspace/vllm/vllm/v1/attention/backends/mla/sparse_swa.py
merged: '2026-09-16'
number: 57152
source:
  - https://github.com/vllm-project/vllm/blob/9f9e1dac26ff0379651dd7e8ca409b573ccfce54/vllm/models/deepseek_v41/common/ops/cache_utils.py
  - https://github.com/vllm-project/vllm/blob/9f9e1dac26ff0379651dd7e8ca409b573ccfce54/vllm/models/deepseek_v41/nvidia/flashmla.py
  - https://github.com/vllm-project/vllm/blob/9f9e1dac26ff0379651dd7e8ca409b573ccfce54/vllm/models/deepseek_v41/nvidia/flash_mla_mega_attn.py
  - https://github.com/vllm-project/vllm/blob/9f9e1dac26ff0379651dd7e8ca409b573ccfce54/vllm/v1/attention/backends/mla/sparse_swa.py
  - https://github.com/vllm-project/vllm/blob/9f9e1dac26ff0379651dd7e8ca409b573ccfce54/vllm/models/deepseek_v41/attention.py
  - https://github.com/vllm-project/vllm/blob/9f9e1dac26ff0379651dd7e8ca409b573ccfce54/vllm/transformers_utils/configs/deepseek_v41.py
  - https://github.com/vllm-project/vllm/blob/9f9e1dac26ff0379651dd7e8ca409b573ccfce54/vllm/models/deepseek_v41/nvidia/flashinfer_sparse.py
state: merged
tags:
  - vllm
  - vllm-pr
  - new-architectures
title: '#57152 [Bugfix][Model] Restore causal image SWA for DeepSeek V4.1'
topic: new-architectures
url: https://github.com/vllm-project/vllm/pull/57152
created: '2026-11-11'
modified: '2026-11-11'
published: '2026-11-11'
pageLayout: default
slug: hinterland/papers/maps/prs/57152
permalink: https://aarnphm.xyz/hinterland/papers/maps/prs/57152.md
generator:
  quartz: v4.6.0
  hostedProvider: Cloudflare
  baseUrl: aarnphm.xyz
full: https://aarnphm.xyz/llms-full.txt
---
Restore causal sliding-window attention for image tokens in DeepSeek V4.1 VL. · full dump: [[thoughts/inference/vllm-prs/57152|#57152]]

