@@ -477,7 +477,7 @@ class Qwen2ModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixi
|
||||
@require_torch_gpu
|
||||
@pytest.mark.flash_attn_test
|
||||
@slow
|
||||
def test_flash_attn_2_inference_padding_right(self):
|
||||
def test_flash_attn_2_inference_equivalence_right_padding(self):
|
||||
self.skipTest("Qwen2 flash attention does not support right padding")
|
||||
|
||||
|
||||
|
||||
Reference in New Issue
Block a user