From 17f05b10893bd18558b3c69f7af880fffc6c1653 Mon Sep 17 00:00:00 2001
From: yiz-liu <136800916+yiz-liu@users.noreply.github.com>
Date: Fri, 23 May 2025 15:50:48 +0800
Subject: [PATCH] [Feature] Add CustomQwen3MoeForCausalLM model (#925)

Tweak packed_modules_mapping to support W8A8 weights.

<!--  Thanks for sending a pull request!

BEFORE SUBMITTING, PLEASE READ
https://docs.vllm.ai/en/latest/contributing/overview.html

-->
### What this PR does / why we need it?
<!--
- Please clarify what changes you are proposing. The purpose of this
section is to outline the changes and how this PR fixes the issue.
If possible, please consider writing useful notes for better and faster
reviews in your PR.

- Please clarify why the changes are needed. For instance, the use case
and bug description.

- Fixes #
-->

### Does this PR introduce _any_ user-facing change?
<!--
Note that it means *any* user-facing change including all aspects such
as API, interface or other behavior changes.
Documentation-only updates are not considered user-facing changes.
-->

### How was this patch tested?
<!--
CI passed with new added/existing test.
If it was tested in a way different from regular unit tests, please
clarify how you tested step by step, ideally copy and paste-able, so
that other reviewers can test and check, and descendants can verify in
the future.
If tests were not added, please describe why they were not added and/or
why it was difficult to add.
-->

Signed-off-by: Yizhou Liu <liu_yizhou@outlook.com>
---
 vllm_ascend/models/__init__.py  |  4 ++++
 vllm_ascend/models/qwen3_moe.py | 35 +++++++++++++++++++++++++++++++++
 2 files changed, 39 insertions(+)
 create mode 100644 vllm_ascend/models/qwen3_moe.py

diff --git a/vllm_ascend/models/__init__.py b/vllm_ascend/models/__init__.py
index 42960f1..e7f021f 100644
--- a/vllm_ascend/models/__init__.py
+++ b/vllm_ascend/models/__init__.py
@@ -29,3 +29,7 @@ def register_model():
     ModelRegistry.register_model(
         "DeepseekV3ForCausalLM",
         "vllm_ascend.models.deepseek_v2:CustomDeepseekV3ForCausalLM")
+
+    ModelRegistry.register_model(
+        "Qwen3MoeForCausalLM",
+        "vllm_ascend.models.qwen3_moe:CustomQwen3MoeForCausalLM")
diff --git a/vllm_ascend/models/qwen3_moe.py b/vllm_ascend/models/qwen3_moe.py
new file mode 100644
index 0000000..8ff1b52
--- /dev/null
+++ b/vllm_ascend/models/qwen3_moe.py
@@ -0,0 +1,35 @@
+# Copyright (c) 2025 Huawei Technologies Co., Ltd. All Rights Reserved.
+# Copyright 2023 The vLLM team.
+#
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# Adapted from vllm/model_executor/models/qwen3_moe.py
+# This file is a part of the vllm-ascend project.
+
+from vllm.model_executor.models.qwen3_moe import Qwen3MoeForCausalLM
+
+
+class CustomQwen3MoeForCausalLM(Qwen3MoeForCausalLM):
+    packed_modules_mapping = {
+        "qkv_proj": [
+            "q_proj",
+            "k_proj",
+            "v_proj",
+        ],
+        "gate_up_proj": [
+            "gate_proj",
+            "up_proj",
+        ],
+        "experts":
+        ["experts.0.gate_proj", "experts.0.up_proj", "experts.0.down_proj"],
+    }