mirror of
https://github.com/NVIDIA/TensorRT-LLM.git
synced 2026-01-14 06:27:45 +08:00
7225bd8b91
Refine attention backend interface. Signed-off-by: Yuxian Qiu <142763828+yuxianq@users.noreply.github.com>
12 lines
650 B
INI
12 lines
650 B
INI
[pytest]
|
|
threadleak = True
|
|
threadleak_exclude = asyncio_\d+
|
|
junit_family=legacy
|
|
addopts = --ignore-glob="*perf/test_perf.py" --ignore-glob="*test_list_validation.py" --ignore-glob="*llm-test-workspace*" --durations=0 -W ignore::DeprecationWarning
|
|
markers =
|
|
skip_less_device: skip when less device detected than the declared
|
|
skip_less_device_memory: skip when less device memory detected than the requested
|
|
skip_less_host_memory: skip when less host memory detected than the requested
|
|
support_fp8: skip when fp8 is not supported on the device
|
|
skip_device_not_contain: skip when the device does not contain the specified keyword
|