[Test] Add tests for the multi-node DeepSeek-V2-Lite network in GE Graph (#4039)

### What this PR does / why we need it?
Add tests for the multi-node DeepSeek-V2-Lite network in GE Graph mode,
and supplement the end-to-end (e2e) tests for the MLA and NZ features of
this network.
- vLLM version: v0.11.0
- vLLM main:
2918c1b49c

---------

Signed-off-by: CodeNine-CJ <chenjian343@huawei.com>
This commit is contained in:
CodeCat
2025-11-20 17:28:32 +08:00
committed by GitHub
parent 3653f33878
commit 470fe05df6
2 changed files with 14 additions and 0 deletions

View File

@@ -171,7 +171,10 @@ jobs:
if: ${{ inputs.type == 'light' }} if: ${{ inputs.type == 'light' }}
run: | run: |
pytest -sv tests/e2e/multicard/test_qwen3_moe.py::test_models_distributed_Qwen3_MOE_TP2_WITH_EP pytest -sv tests/e2e/multicard/test_qwen3_moe.py::test_models_distributed_Qwen3_MOE_TP2_WITH_EP
pytest -sv tests/e2e/multicard/test_torchair_graph_mode.py::test_e2e_qwen3_moe_with_torchair
pytest -sv tests/e2e/multicard/test_torchair_graph_mode.py::test_e2e_deepseekv2lite_with_torchair pytest -sv tests/e2e/multicard/test_torchair_graph_mode.py::test_e2e_deepseekv2lite_with_torchair
pytest -sv tests/e2e/multicard/test_torchair_graph_mode.py::test_e2e_deepseekv2lite_with_torchair_v1scheduler
pytest -sv tests/e2e/multicard/test_torchair_graph_mode.py::test_e2e_deepseekv2lite_with_nz
- name: Run vllm-project/vllm-ascend test (full) - name: Run vllm-project/vllm-ascend test (full)
env: env:

View File

@@ -288,3 +288,14 @@ def test_e2e_deepseekv2lite_with_torchair_v1scheduler():
} }
_deepseek_v2_lite_torchair_test_fixure(additional_config, _deepseek_v2_lite_torchair_test_fixure(additional_config,
use_v1_schduler=True) use_v1_schduler=True)
# kv_cache enable e2e test
def test_e2e_deepseekv2lite_with_nz():
additional_config = {
"torchair_graph_config": {
"enabled": True,
"enable_kv_nz": True,
},
}
_deepseek_v2_lite_torchair_test_fixure(additional_config)