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

[Frontend] Support using chat template as custom score template for reranking models (#30550)

Browse Source 
Signed-off-by: Jakub Zakrzewski <jzakrzewski@nvidia.com>
Signed-off-by: wang.yuqi <yuqi.wang@daocloud.io>
Signed-off-by: wang.yuqi <noooop@126.com>
Co-authored-by: wang.yuqi <yuqi.wang@daocloud.io>
This commit is contained in:
Jakub Zakrzewski
2025-12-23 12:19:16 +01:00
committed by GitHub
parent 27c6c2f98c
commit 23daef548d
19 changed files with 663 additions and 46 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
vllm/model_executor/models/registry.py
View File

@@ -203,6 +203,7 @@ _EMBEDDING_MODELS = {
"GteNewModel": ("bert_with_rope", "GteNewModel"),
"InternLM2ForRewardModel": ("internlm2", "InternLM2ForRewardModel"),
"JambaForSequenceClassification": ("jamba", "JambaForSequenceClassification"), # noqa: E501
"LlamaBidirectionalModel": ("llama", "LlamaBidirectionalModel"),
"LlamaModel": ("llama", "LlamaForCausalLM"),
**{
# Multiple models share the same architecture, so we include them all
@@ -246,6 +247,11 @@ _CROSS_ENCODER_MODELS = {
"bert_with_rope",
"GteNewForSequenceClassification",
),
"JinaVLForRanking": ("jina_vl", "JinaVLForSequenceClassification"),
"LlamaBidirectionalForSequenceClassification": (
"llama",
"LlamaBidirectionalForSequenceClassification",
),
"ModernBertForSequenceClassification": (
"modernbert",
"ModernBertForSequenceClassification",
@@ -259,8 +265,6 @@ _CROSS_ENCODER_MODELS = {
"roberta",
"RobertaForSequenceClassification",
),
# [Auto-converted (see adapters.py)]
"JinaVLForRanking": ("jina_vl", "JinaVLForSequenceClassification"), # noqa: E501,
}
_MULTIMODAL_MODELS = {