---
cited_by:
  - '[[thoughts/inference/qwen3.8-serving#4-pr-reading-order-vllm-main-git-show-sha|qwen4next §4]]'
date: '2026-11-11'
kind: pr
local:
  - file:///Users/aarnphm/workspace/vllm/vllm/v1/attention/backends/short_conv_attn.py
merged: '2026-09-02'
number: 54517
source:
  - https://github.com/vllm-project/vllm/blob/f870b9297685bd3c968063b52b769315cb08fd7f/vllm/models/qwen4_exp/nvidia/ple_layer.py
  - https://github.com/vllm-project/vllm/blob/f870b9297685bd3c968063b52b769315cb08fd7f/vllm/models/qwen4_exp/nvidia/ops/ple.py
  - https://github.com/vllm-project/vllm/blob/f870b9297685bd3c968063b52b769315cb08fd7f/vllm/models/qwen4_exp/nvidia/model.py
  - https://github.com/vllm-project/vllm/blob/f870b9297685bd3c968063b52b769315cb08fd7f/vllm/v1/attention/backends/short_conv_attn.py
  - https://github.com/vllm-project/vllm/blob/f870b9297685bd3c968063b52b769315cb08fd7f/vllm/models/qwen4_exp/nvidia/mtp.py
state: merged
tags:
  - vllm
  - vllm-pr
  - new-architectures
title: '#54517 [Qwen3.8-Flash-Next] Fuse Qwen4Exp PLE kernels'
topic: new-architectures
url: https://github.com/vllm-project/vllm/pull/54517
created: '2026-11-11'
modified: '2026-11-11'
published: '2026-11-11'
pageLayout: default
slug: hinterland/papers/maps/prs/54517
permalink: https://aarnphm.xyz/hinterland/papers/maps/prs/54517.md
generator:
  quartz: v4.6.0
  hostedProvider: Cloudflare
  baseUrl: aarnphm.xyz
full: https://aarnphm.xyz/llms-full.txt
---
Fuse n-gram ID generation, dilated short-conv, and the PLE gate on NVIDIA · full dump: [[thoughts/inference/vllm-prs/54517|#54517]]

