[Misc] Split up pooling tasks (#10820)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>

tests/models/embedding/vision_language/test_dse_qwen2_vl.py

@@ -93,7 +93,7 @@ def _run_test(
     # if we run HF first, the cuda initialization will be done and it
     # will hurt multiprocessing backend with fork method (the default method).
     with vllm_runner(model,
-                     task="embedding",
+                     task="embed",
                      dtype=dtype,
                      enforce_eager=True,
                      max_model_len=8192) as vllm_model:

tests/models/embedding/vision_language/test_llava_next.py

@@ -47,7 +47,7 @@ def _run_test(
     # if we run HF first, the cuda initialization will be done and it
     # will hurt multiprocessing backend with fork method (the default method).
     with vllm_runner(model,
-                     task="embedding",
+                     task="embed",
                      dtype=dtype,
                      max_model_len=4096,
                      enforce_eager=True) as vllm_model:

tests/models/embedding/vision_language/test_phi3v.py

@@ -39,7 +39,7 @@ def _run_test(
     # vLLM needs a fresh new process without cuda initialization.
     # if we run HF first, the cuda initialization will be done and it
     # will hurt multiprocessing backend with fork method (the default method).
-    with vllm_runner(model, task="embedding", dtype=dtype,
+    with vllm_runner(model, task="embed", dtype=dtype,
                      enforce_eager=True) as vllm_model:
         vllm_outputs = vllm_model.encode(input_texts, images=input_images)