skip flash attn ut on Hopper

[Wong4j]

PR types

Others

PR changes

Others

Description

The test_flash_attention.py UT failed on H100 because flash attention is only implemented for sm75 and sm8x. Refer to flash_attn.cpp#L293

So I skip this UT on unsupported arch.

Error message:

ERROR: test_all (test_flash_attention.TestFlashAttentionAPI)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/opt/paddle/paddle/build/python/paddle/fluid/tests/unittests/test_flash_attention.py", line 192, in test_all
    out, _ = flash_attention(
  File "/opt/paddle/paddle/build/python/paddle/nn/functional/flash_attention.py", line 83, in flash_attention
    (result_attention, result_softmax,) = _C_ops.flash_attn(
OSError: (External) `is_sm8x || is_sm75` check failed at /opt/paddle/paddle/build/third_party/flashattn/src/extern_flashattn/csrc/flash_attn/flash_attn.cpp:292 (at /opt/paddle/paddle/paddle/phi/kernels/gpu/flash_attn_kernel.cu:142)

[paddle-bot]

你的PR提交成功，感谢你对开源项目的贡献!
请关注后续CI自动化测试结果，详情请参考Paddle-CI手册。
Your PR has been submitted. Thanks for your contribution!
Please wait for the result of CI firstly. See Paddle CI Manual for details.

[XieYunshen]

LGTM

[paddle-bot]

你的PR已合入Paddle库，请关注后续测试结果。
Your PR has been merged into the repository. An official integration test will be conducted later. Stay tuned.