EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

[DeepSeek] Seperate deepseek v3.2 modeling form deepseek v2 (#3531)

Browse Source 
### What this PR does / why we need it?
Seperate deepseek v3.2 modeling form deepseek v2

### How was this patch tested?
- CI passed with existing test.
- test deepseek v3.2 locally

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

Signed-off-by: MengqingCao <cmq0113@163.com>
This commit is contained in:
Mengqing Cao
2025-10-20 09:50:44 +08:00
committed by GitHub
parent 6c65dd891f
commit daa4dd0a57
4 changed files with 637 additions and 193 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
vllm_ascend/models/__init__.py
View File

@@ -39,7 +39,7 @@ def register_model():
ModelRegistry.register_model(
"DeepseekV32ForCausalLM",
"vllm_ascend.models.deepseek_v2:CustomDeepseekV3ForCausalLM")
"vllm_ascend.models.deepseek_v3_2:CustomDeepseekV3ForCausalLM")
ModelRegistry.register_model(
"DeepSeekMTPModel",