A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 13, 14:30 UTC
[CI] lora/test_mixtral.py : Add additional expected outputs due to flakiness (#28322)
Signed-off-by: Varun Sundar Rabindranath <vsundarr@redhat.com> Co-authored-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
Varun Sundar Rabindranath committed
Nov 9, 2025 at 21:45 UTC
6b2b9fd934af2bf59f749aa04ac739fd9c7d29f7
1 file changed
+18
-11
tests/lora/test_mixtral.py
+18
-11
@@ -56,15 +56,22 @@ def test_mixtral_lora(mixtral_lora_files, tp_size):
56
)
57
58
expected_lora_output = [
59
- "give_opinion(name[SpellForce 3], release_year[2017], developer[Grimlore Games], rating[poor])", # noqa: E501
60
- "give_opinion(name[SpellForce 3], developer[Grimlore Games], release_year[2017], rating[poor])", # noqa: E501
61
- "inform(name[BioShock], release_year[2007], rating[good], genres[action-adventure, role-playing, shooter], platforms[PlayStation, Xbox, PC], available_on_steam[yes], has_linux_release[no], has_mac_release[yes])", # noqa: E501
59
+ [
60
+ "give_opinion(name[SpellForce 3], release_year[2017], developer[Grimlore Games], rating[poor])" # noqa: E501
61
+ ],
62
+ [
63
+ "give_opinion(name[SpellForce 3], developer[Grimlore Games], release_year[2017], rating[poor])", # noqa: E501
64
+ "give_opinion(name[SpellForce 3], release_year[2017], developer[Grimlore Games], rating[poor])", # noqa: E501
65
+ ],
66
+ [
67
+ "inform(name[BioShock], release_year[2007], rating[good], genres[action-adventure, role-playing, shooter], platforms[PlayStation, Xbox, PC], available_on_steam[yes], has_linux_release[no], has_mac_release[yes])" # noqa: E501
68
+ ],
69
]
63
- assert (
64
- do_sample(llm, mixtral_lora_files, lora_id=1, prompts=prompts)
65
- == expected_lora_output
66
- )
67
- assert (
68
- do_sample(llm, mixtral_lora_files, lora_id=2, prompts=prompts)
69
- == expected_lora_output
70
- )
70
+
71
+ def check_outputs(generated: list[str]):
72
+ assert len(generated) == len(expected_lora_output)
73
+ for gen, gt_choices in zip(generated, expected_lora_output):
74
+ assert gen in gt_choices
75
+
76
+ check_outputs(do_sample(llm, mixtral_lora_files, lora_id=1, prompts=prompts))
77
+ check_outputs(do_sample(llm, mixtral_lora_files, lora_id=2, prompts=prompts))