[Model] Deprecate the score task (this will not affect users). (#37537)

Signed-off-by: wang.yuqi <yuqi.wang@daocloud.io>
This commit is contained in:
wang.yuqi
2026-03-20 08:07:56 +00:00
committed by GitHub
parent dcee9be95a
commit ed359c497a
22 changed files with 184 additions and 163 deletions
@@ -194,18 +194,18 @@ class VllmModelForPooling(VllmModel[T_co], Protocol[T_co]):
[vllm.config.model.ModelConfig.score_type][]
to use by default.
Score API handles score/rerank for:
- "score" task (score_type: cross-encoder models)
- "embed" task (score_type: bi-encoder models)
- "token_embed" task (score_type: late interaction models)
Scoring API handles score/rerank for:\n
- "classify" task (score_type: cross-encoder models)\n
- "embed" task (score_type: bi-encoder models)\n
- "token_embed" task (score_type: late interaction models)\n
score_type defaults to bi-encoder, then the Score API uses the "embed" task.
score_type defaults to bi-encoder, then the Score API uses the "embed" task.\n
If you set score_type to cross-encoder via
[vllm.model_executor.models.interfaces.SupportsCrossEncoding][],
then the Score API uses the "score" task.
then the Score API uses the "score" task.\n
If you set score_type to late-interaction via
[vllm.model_executor.models.interfaces.SupportsLateInteraction][],
then the Score API uses the "token_embed" task.
then the Score API uses the "token_embed" task.\n
"""
pooler: Pooler