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

[CI] Reorganization pooling_mteb_test (#31265)

Browse Source 
Signed-off-by: wang.yuqi <noooop@126.com>
This commit is contained in:
wang.yuqi
2025-12-24 23:36:20 +08:00
committed by GitHub
parent 7cd288a4b3
commit 1ff67df182
18 changed files with 480 additions and 347 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
tests/models/language/pooling_mteb_test/test_nemotron.py
View File

@@ -3,29 +3,39 @@
import pytest
from tests.models.language.pooling_mteb_test.mteb_embed_utils import (
mteb_test_embed_models,
)
from tests.models.language.pooling_mteb_test.mteb_score_utils import (
mteb_test_rerank_models,
)
from tests.models.utils import (
EmbedModelInfo,
LASTPoolingEmbedModelInfo,
LASTPoolingRerankModelInfo,
RerankModelInfo,
)
from .mteb_utils import mteb_test_embed_models, mteb_test_rerank_models
EMBEDDING_MODELS = [
LASTPoolingEmbedModelInfo(
EmbedModelInfo(
"nvidia/llama-nemotron-embed-1b-v2",
architecture="LlamaBidirectionalModel",
mteb_score=0.689164662128673,
pooling_type="MEAN",
attn_type="encoder_only",
is_prefix_caching_supported=False,
is_chunked_prefill_supported=False,
)
]
RERANK_MODELS = [
LASTPoolingRerankModelInfo(
RerankModelInfo(
"nvidia/llama-nemotron-rerank-1b-v2",
architecture="LlamaBidirectionalForSequenceClassification",
chat_template_name="nemotron-rerank.jinja",
mteb_score=0.33994,
pooling_type="MEAN",
attn_type="encoder_only",
is_prefix_caching_supported=False,
is_chunked_prefill_supported=False,
),
]