Skip to content

[https://nvbugs/5451028][fix] Constrain NemotronSuper test parameters… #6987

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Naveassaf merged 1 commit into from
Aug 19, 2025
Merged

[https://nvbugs/5451028][fix] Constrain NemotronSuper test parameters… #6987

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 11 additions & 12 deletions tests/integration/defs/accuracy/test_llm_api_pytorch.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1747,29 +1747,28 @@ class TestLlama3_3NemotronSuper49Bv1(LlmapiAccuracyTestHarness):
@pytest.mark.skip_less_device(2)
@pytest.mark.skip_less_device_memory(80000)
def test_auto_dtype_tp2(self):
with LLM(self.MODEL_PATH, tensor_parallel_size=2) as llm:
with LLM(self.MODEL_PATH,
tensor_parallel_size=2,
max_seq_len=8192,
max_batch_size=64) as llm:
# Run only one eval as maximal BS is not large
task = MMLU(self.MODEL_NAME)
task.evaluate(llm)
task = GSM8K(self.MODEL_NAME)
task.evaluate(llm)
task = GPQADiamond(self.MODEL_NAME)
task.evaluate(llm,
extra_evaluator_kwargs=dict(apply_chat_template=True))

@skip_pre_hopper
@pytest.mark.skip_less_device(2)
@pytest.mark.skip_device_not_contain(["H100", "B200"])
def test_fp8_prequantized_tp2(self):
model_path = f"{llm_models_root()}/nemotron-nas/Llama-3_3-Nemotron-Super-49B-v1-FP8"
with LLM(model_path, tensor_parallel_size=2) as llm:
with LLM(model_path,
tensor_parallel_size=2,
max_seq_len=8192,
max_batch_size=64) as llm:
assert llm.args.quant_config.quant_algo == QuantAlgo.FP8

# Run only one eval as maximal BS is not large
task = MMLU(self.MODEL_NAME)
task.evaluate(llm)
task = GSM8K(self.MODEL_NAME)
task.evaluate(llm)
task = GPQADiamond(self.MODEL_NAME)
task.evaluate(llm,
extra_evaluator_kwargs=dict(apply_chat_template=True))


class TestLlama3_1NemotronNano8Bv1(LlmapiAccuracyTestHarness):
Expand Down