[Core] Deprecate xformers (#29262)

Signed-off-by: Roger Wang <hey@rogerw.io>
This commit is contained in:
Roger Wang
2025-11-23 20:18:55 -08:00
committed by GitHub
parent 5253f4276f
commit 0ff70821c9
31 changed files with 77 additions and 963 deletions

View File

@@ -57,10 +57,6 @@ def do_sample(llm: vllm.LLM, lora_path: str, lora_id: int) -> list[str]:
return generated_texts
@pytest.mark.xfail(
current_platform.is_rocm(),
reason="MiniCPM-V dependency xformers incompatible with ROCm",
)
def test_minicpmv_lora(minicpmv_lora_files):
llm = vllm.LLM(
MODEL_PATH,
@@ -84,10 +80,6 @@ def test_minicpmv_lora(minicpmv_lora_files):
@pytest.mark.skipif(
current_platform.is_cuda_alike(), reason="Skipping to avoid redundant model tests"
)
@pytest.mark.xfail(
current_platform.is_rocm(),
reason="MiniCPM-V dependency xformers incompatible with ROCm",
)
@multi_gpu_test(num_gpus=4)
def test_minicpmv_tp4_wo_fully_sharded_loras(minicpmv_lora_files):
llm = vllm.LLM(
@@ -108,10 +100,6 @@ def test_minicpmv_tp4_wo_fully_sharded_loras(minicpmv_lora_files):
@pytest.mark.skipif(
current_platform.is_cuda_alike(), reason="Skipping to avoid redundant model tests"
)
@pytest.mark.xfail(
current_platform.is_rocm(),
reason="MiniCPM-V dependency xformers incompatible with ROCm",
)
@multi_gpu_test(num_gpus=4)
def test_minicpmv_tp4_fully_sharded_loras(minicpmv_lora_files):
llm = vllm.LLM(