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

[CI] Add DeepSeek-V3.2-W8A8 nightly ci test (#4633)

Browse Source 
### What this PR does / why we need it?
Add DeepSeek-V3.2-W8A8 nightly ci test：

DeepSeek-V3.2-W8A8 1node DP2+TP8
:tests/e2e/nightly/models/test_deepseek_v3_2_w8a8.py

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

- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c

Signed-off-by: hfadzxy <starmoon_zhang@163.com>
This commit is contained in:
zhangxinyuehfad
2026-01-20 21:05:15 +08:00
committed by GitHub
parent cea48c2a34
commit 750c06c78a
6 changed files with 30 additions and 93 deletions
Download Patch File Download Diff File Expand all files Collapse all files

51
tests/e2e/nightly/multi_node/config/DeepSeek-V3_2-Exp-bf16.yaml
View File

@@ -1,51 +0,0 @@
test_name: "test DeepSeek-V3.2-Exp-bf16 multi-dp"
model: "Yanguan/DeepSeek-V3.2-Exp-bf16"
num_nodes: 2
npu_per_node: 16
env_common:
VLLM_USE_MODELSCOPE: true
OMP_PROC_BIND: false
OMP_NUM_THREADS: 100
HCCL_BUFFSIZE: 1024
SERVER_PORT: 8080
VLLM_ASCEND_ENABLE_MLAPO: 0
deployment:
-
server_cmd: >
vllm serve "Yanguan/DeepSeek-V3.2-Exp-bf16"
--host 0.0.0.0
--port $SERVER_PORT
--data-parallel-address $LOCAL_IP
--data-parallel-size 2
--data-parallel-size-local 1
--data-parallel-rpc-port 13389
--tensor-parallel-size 16
--seed 1024
--enable-expert-parallel
--max-num-seqs 16
--max-model-len 17450
--max-num-batched-tokens 17450
--trust-remote-code
--no-enable-prefix-caching
--gpu-memory-utilization 0.9
-
server_cmd: >
vllm serve "Yanguan/DeepSeek-V3.2-Exp-bf16"
--headless
--data-parallel-size 2
--data-parallel-size-local 1
--data-parallel-start-rank 1
--data-parallel-address $MASTER_IP
--data-parallel-rpc-port 13389
--tensor-parallel-size 16
--seed 1024
--max-num-seqs 16
--max-model-len 17450
--max-num-batched-tokens 17450
--enable-expert-parallel
--trust-remote-code
--no-enable-prefix-caching
--gpu-memory-utilization 0.92
benchmarks: