[v0.11.0-dev][Bugfix][cherry-pick]bugfix for weight load of kimi-k2 (#4190)

### What this PR does / why we need it?
This is cherry-pick from #3798 

Fix kimi-k2 start bug, weight load
ERROR:https://github.com/vllm-project/vllm-ascend/issues/3785

### Does this PR introduce _any_ user-facing change?

### How was this patch tested?

- vLLM version: v0.11.0rc3
- vLLM main:
https://github.com/vllm-project/vllm/commit/c9461e05a4ed3557cfbf4b15ded1e26761cc39ca

---------

Signed-off-by: Levi-JQ <yujinqi2@huawei.com>
Signed-off-by: menogrey <1299267905@qq.com>
Co-authored-by: Levi <54832289+Levi-JQ@users.noreply.github.com>
Co-authored-by: Levi-JQ <yujinqi2@huawei.com>
Co-authored-by: zhaozx-cn <zhaozx2116@163.com>
This commit is contained in:
authored and GitHub committed 2025-11-14 15:43:22 +08:00
1 parent 51e5806d76
commit a7eb42cf0a
2 files changed
+12 -1

No files matched your search

+7 -1
View File
@@ -57,7 +57,13 @@ jobs:
- name: Print
run: |
lscpu
- name: Free up disk space
uses: jlumbroso/free-disk-space@54081f138730dfa15788a46383842cd2f914a1be # v1.3.1
with:
tool-cache: true
docker-images: false
- name: Build wheel
run: |
ls
+5
View File
@@ -193,6 +193,11 @@ packed_modules_model_mapping = {
["experts.0.gate_proj", "experts.0.up_proj", "experts.0.down_proj"],
"fused_qkv_a_proj": ["q_a_proj", "kv_a_proj_with_mqa"]
},
"kimi_k2": {
"gate_up_proj": ["gate_proj", "up_proj"],
"experts":
["experts.0.gate_proj", "experts.0.up_proj", "experts.0.down_proj"]
},
"deepseek_v32": {
"gate_up_proj": ["gate_proj", "up_proj"],
"experts":