biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Config] add "qwen" as a native eagle3 target supported model (#22333)

Browse Source 
Signed-off-by: lechen <lecself@163.com>
Signed-off-by: LeChen <lecself@163.com>
This commit is contained in:
Le Chen
2025-08-10 11:21:05 +08:00
committed by GitHub
parent 0c5254b82a
commit 3d7363e61c
5 changed files with 30 additions and 27 deletions
Download Patch File Download Diff File Expand all files Collapse all files

39
tests/v1/e2e/test_spec_decode.py
View File

@@ -125,24 +125,27 @@ def test_ngram_correctness(
cleanup_dist_env_and_memory()
@pytest.mark.parametrize(
["model_setup", "mm_enabled"], [
(("eagle", "meta-llama/Llama-3.1-8B-Instruct",
"yuhuili/EAGLE-LLaMA3.1-Instruct-8B", 1), False),
(("eagle3", "meta-llama/Llama-3.1-8B-Instruct",
"yuhuili/EAGLE3-LLaMA3.1-Instruct-8B", 1), False),
pytest.param(
("eagle", "meta-llama/Llama-4-Scout-17B-16E-Instruct",
"morgendave/EAGLE-Llama-4-Scout-17B-16E-Instruct", 4),
False,
marks=pytest.mark.skip(reason="Skipping due to CI OOM issues")),
pytest.param(
("eagle", "meta-llama/Llama-4-Scout-17B-16E-Instruct",
"morgendave/EAGLE-Llama-4-Scout-17B-16E-Instruct", 4),
True,
marks=pytest.mark.skip(reason="Skipping due to CI OOM issues")),
],
ids=["llama3_eagle", "llama3_eagle3", "llama4_eagle", "llama4_eagle_mm"])
@pytest.mark.parametrize(["model_setup", "mm_enabled"], [
(("eagle3", "Qwen/Qwen3-8B", "AngelSlim/Qwen3-8B_eagle3", 1), False),
(("eagle", "meta-llama/Llama-3.1-8B-Instruct",
"yuhuili/EAGLE-LLaMA3.1-Instruct-8B", 1), False),
(("eagle3", "meta-llama/Llama-3.1-8B-Instruct",
"yuhuili/EAGLE3-LLaMA3.1-Instruct-8B", 1), False),
pytest.param(
("eagle", "meta-llama/Llama-4-Scout-17B-16E-Instruct",
"morgendave/EAGLE-Llama-4-Scout-17B-16E-Instruct", 4),
False,
marks=pytest.mark.skip(reason="Skipping due to CI OOM issues")),
pytest.param(
("eagle", "meta-llama/Llama-4-Scout-17B-16E-Instruct",
"morgendave/EAGLE-Llama-4-Scout-17B-16E-Instruct", 4),
True,
marks=pytest.mark.skip(reason="Skipping due to CI OOM issues")),
],
ids=[
"qwen3_eagle3", "llama3_eagle", "llama3_eagle3",
"llama4_eagle", "llama4_eagle_mm"
])
@pytest.mark.parametrize("attn_backend",
get_attn_backend_list_based_on_platform())
def test_eagle_correctness(