Skip to content

[https://nvbugs/5410687][test] Add deepseek r1-w4afp8 quickstart #7645

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
crazydemo merged 2 commits into from
Sep 10, 2025
Merged

[https://nvbugs/5410687][test] Add deepseek r1-w4afp8 quickstart #7645

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
61bd0d4
feature: add rcca deepseekr1 int4 afp8
fredricz-20070104 Sep 9, 2025
aca70f7
merge upstream main branch
fredricz-20070104 Sep 9, 2025
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
38 changes: 36 additions & 2 deletions tests/integration/defs/test_e2e.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,8 +32,8 @@
prune_checkpoint, quantize_data, refit_model,
venv_check_call)
from .conftest import (llm_models_root, skip_no_sm120, skip_nvlink_inactive,
skip_post_blackwell, skip_pre_blackwell, skip_pre_hopper,
tests_path, unittest_path)
skip_post_blackwell, skip_pre_ada, skip_pre_blackwell,
skip_pre_hopper, tests_path, unittest_path)

sys.path.append(os.path.join(str(tests_path()), '/../examples/apps'))

Expand Down Expand Up @@ -2155,6 +2155,40 @@ def test_relaxed_acceptance_quickstart_advanced_deepseek_r1_8gpus(
_check_mem_usage(running_log, [85.6, 0, 0, 0], 8)


@skip_pre_ada
@skip_post_blackwell
@pytest.mark.skip_less_device_memory(80000)
@pytest.mark.skip_less_device(8)
@pytest.mark.parametrize("model_name,model_path", [
pytest.param('DeepSeek-R1-W4AFP8',
'DeepSeek-R1/DeepSeek-R1-W4AFP8',
marks=skip_pre_hopper),
])
def test_ptp_quickstart_advanced_deepseek_r1_w4afp8_8gpus(
llm_root, llm_venv, model_name, model_path):
print(f"Testing {model_name}.")
example_root = Path(os.path.join(llm_root, "examples", "llm-api"))
with tempfile.NamedTemporaryFile(mode='w+t',
suffix=f".{model_name}.log",
dir="./",
delete=True,
delete_on_close=True) as running_log:
llm_venv.run_cmd([
str(example_root / "quickstart_advanced.py"),
"--model_dir",
f"{llm_models_root()}/{model_path}",
"--moe_tp_size=1",
"--moe_ep_size=8",
"--tp_size=8",
"--use_cuda_graph",
f"--kv_cache_fraction={_MEM_FRACTION_50}",
"--max_batch_size=1",
"--max_seq_len=512",
],
stdout=running_log)
_check_mem_usage(running_log, [50.0, 0, 0, 0], 8)


@pytest.mark.skip_less_device_memory(80000)
@pytest.mark.skip_less_device(8)
@pytest.mark.parametrize("model_name,model_path", [
Expand Down
1 change: 1 addition & 0 deletions tests/integration/test_lists/qa/llm_function_core.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -645,6 +645,7 @@ test_e2e.py::test_ptp_quickstart_bert[TRTLLM-BertForSequenceClassification-bert/
test_e2e.py::test_ptp_star_attention_example[Llama3.1-8B-BF16-llama-3.1-model/Meta-Llama-3.1-8B]
test_e2e.py::test_trtllm_bench_pytorch_backend_sanity[meta-llama/Llama-3.1-8B-llama-3.1-8b-hf-nvfp4-False-False]
test_e2e.py::test_ptp_scaffolding[DeepSeek-R1-Distill-Qwen-7B-DeepSeek-R1/DeepSeek-R1-Distill-Qwen-7B]
test_e2e.py::test_ptp_quickstart_advanced_deepseek_r1_w4afp8_8gpus[DeepSeek-R1-W4AFP8-DeepSeek-R1/DeepSeek-R1-W4AFP8]
Comment thread
fredricz-20070104 marked this conversation as resolved.
unittest/llmapi/test_llm_pytorch.py::test_gemma3_1b_instruct_multi_lora
examples/test_medusa.py::test_codellama_medusa_1gpu[CodeLlama-7b-Instruct]

Expand Down
1 change: 1 addition & 0 deletions tests/integration/test_lists/qa/llm_function_core_sanity.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -215,6 +215,7 @@ test_e2e.py::test_ptp_quickstart_multimodal[qwen2.5-vl-7b-instruct-Qwen2.5-VL-7B
test_e2e.py::test_ptp_quickstart_multimodal[qwen2.5-vl-7b-instruct-Qwen2.5-VL-7B-Instruct-video-False]
test_e2e.py::test_ptp_quickstart_multimodal[qwen2.5-vl-7b-instruct-Qwen2.5-VL-7B-Instruct-video-True]
test_e2e.py::test_ptp_scaffolding[DeepSeek-R1-Distill-Qwen-7B-DeepSeek-R1/DeepSeek-R1-Distill-Qwen-7B]
test_e2e.py::test_ptp_quickstart_advanced_deepseek_r1_w4afp8_8gpus[DeepSeek-R1-W4AFP8-DeepSeek-R1/DeepSeek-R1-W4AFP8]
test_e2e.py::test_ptp_star_attention_example[Llama3.1-8B-BF16-llama-3.1-model/Meta-Llama-3.1-8B]
test_e2e.py::test_qwen_e2e_cpprunner_large_new_tokens[DeepSeek-R1-Distill-Qwen-1.5B-DeepSeek-R1-Distill-Qwen-1.5B]
test_e2e.py::test_relaxed_acceptance_quickstart_advanced_deepseek_r1_8gpus[DeepSeek-R1-DeepSeek-R1/DeepSeek-R1]
Expand Down
Loading