[VLMs] fix flash-attention tests (#37603)

* fix one test

* fa2 ln test

* remove keys from config recursively

* fix

* fixup
This commit is contained in:
Raushan Turganbay
2025-04-24 11:48:11 +02:00
committed by GitHub
parent 02baa61fab
commit 1cfcbfcab8
17 changed files with 52 additions and 83 deletions

View File

@@ -367,10 +367,6 @@ class Idefics3ForConditionalGenerationModelTest(GenerationTesterMixin, ModelTest
def test_prompt_lookup_decoding_matches_greedy_search(self):
pass
@unittest.skip(reason=" FlashAttention only support fp16 and bf16 data type")
def test_flash_attn_2_fp32_ln(self):
pass
@pytest.mark.generate
@require_torch_sdpa
@slow