[Model][1/N] Support multiple poolers at model level (#21227)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-07-21 17:22:21 +08:00
parent 378d33c392
commit 042af0c8d3
22 changed files with 549 additions and 413 deletions
--- a/tests/models/test_transformers.py
+++ b/tests/models/test_transformers.py
@@ -144,7 +144,7 @@ def test_quantization(
    "model",
    ["jason9693/Qwen2.5-1.5B-apeach"],
 )
-@pytest.mark.parametrize("dtype", ["half"])
+@pytest.mark.parametrize("dtype", ["float"])
 def test_classify(
    hf_runner,
    vllm_runner,