test template

.github/workflows/cpu-test.yml

@@ -32,3 +32,7 @@ jobs:
     - name: cpu-test
       run: |
         VLLM_SKIP_WARMUP=true VLLM_PROMPT_SEQ_BUCKET_MAX=128 python examples/offline_inference_fakehpu.py
+    - name: pytest-test
+      run: |
+        pytest tests/hpu/
+    

tests/hpu/test_habana_attention.py

@@ -0,0 +1,8 @@
+import vllm.attention.backends.habana_attn 
+import vllm.attention.ops.habana_paged_attn
+
+def test_habana_attn():
+    assert True == True
+
+def test_habana_paged_attn():
+    assert True == True

tests/hpu/test_habana_executor.py

@@ -0,0 +1,8 @@
+from vllm.executor.habana_executor import HabanaExecutor
+from vllm.executor.ray_habana_executor import RayHabanaExecutor
+
+def test_habana_executor():
+    assert True == True
+
+def test_ray_habana_executor():
+    assert True == True

tests/hpu/test_habana_worker.py

@@ -0,0 +1,8 @@
+from vllm.worker.habana_model_runner import HabanaModelRunner
+from vllm.worker.habana_worker import HabanaWorker
+
+def test_habana_worker():
+    assert True == True
+
+def test_habana_model_runner():
+    assert True == True