vllm/tests/basic_correctness/test_cpu_offload.py

from ..utils import compare_two_settings


def test_cpu_offload():
    compare_two_settings("meta-llama/Llama-3.2-1B", [],
                         ["--cpu-offload-gb", "1"])
[ci][test] add correctness test for cpu offloading (#6549) 2024-07-18 16:41:06 -07:00			`from ..utils import compare_two_settings`


			`def test_cpu_offload():`
[CI/Build] Replaced some models on tests for smaller ones (#9570) Signed-off-by: Wallas Santos <wallashss@ibm.com> 2024-10-22 01:52:14 -03:00			`compare_two_settings("meta-llama/Llama-3.2-1B", [],`
			`["--cpu-offload-gb", "1"])`