File tree Expand file tree Collapse file tree 1 file changed +5
-4
lines changed
tests/models/quantization Expand file tree Collapse file tree 1 file changed +5
-4
lines changed Original file line number Diff line number Diff line change 1010
1111from tests .quantization .utils import is_quant_method_supported
1212from vllm .platforms import current_platform
13- from vllm .platforms .rocm import on_gfx9
1413
1514from ...utils import compare_two_settings , multi_gpu_test
1615from ..utils import check_embeddings_close , check_logprobs_close
1716
18- pytestmark = pytest .mark .skipif (
19- current_platform .is_rocm () and on_gfx9 (),
20- reason = "bitsandbytes quantization not supported on gfx9 (warp size 64 limitation)" ,
17+ if current_platform .is_rocm ():
18+ from vllm .platforms .rocm import on_gfx9
19+ pytestmark = pytest .mark .skipif (
20+ on_gfx9 (),
21+ reason = "bitsandbytes quantization not supported on gfx9 (warp size 64 limitation)" ,
2122)
2223
2324models_4bit_to_test = [
You can’t perform that action at this time.
0 commit comments