minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

[MoE Refactor] [NPU] Refactor Ascend MoE implementation to reduce code duplication and align with community design (#25663)

Browse Source 
Co-authored-by: ronnie_zheng <zl19940307@163.com>
This commit is contained in:
Артем Савкин
2026-07-15 14:59:42 +03:00
committed by GitHub
co-authored by ronnie_zheng
parent c9b17403e7
commit 8ed82afcc8
54 changed files with 2851 additions and 2375 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs_new/docs/references/environment_variables.mdx
-5
View File
@@ -1831,11 +1831,6 @@ SGLang supports various environment variables that can be used to configure its
<td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.05)"}}>Use the MLAPO path on NPU.</td>
<td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.02)"}}><code>false</code></td>
</tr>
<tr>
<td style={{padding: "9px 12px", fontWeight: 500, backgroundColor: "rgba(255,255,255,0.02)"}}><code>SGLANG_NPU_FUSED_MOE_MODE</code></td>
<td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.05)"}}>Fused MoE mode selector for NPU.</td>
<td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.02)"}}><code>1</code></td>
</tr>
<tr>
<td style={{padding: "9px 12px", fontWeight: 500, backgroundColor: "rgba(255,255,255,0.02)"}}><code>SGLANG_NPU_FORWARD_NATIVE_GELUTANH</code></td>
<td style={{padding: "9px 12px", backgroundColor: "rgba(255,255,255,0.05)"}}>Use the native gelu-tanh activation forward (for Skywork-Reward-Gemma-2-27B-v0.2).</td>