[jira] [Closed] (OPENNLP-1977) SentenceVectorsDL returns the [CLS] hidden state instead of the pooled sentence embedding

["Richard Zowalla (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/OPENNLP-1977?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Richard Zowalla closed OPENNLP-1977.
------------------------------------
    Resolution: Fixed

Fixed by https://github.com/apache/opennlp/pull/1310 (6331d6e19).

> SentenceVectorsDL returns the [CLS] hidden state instead of the pooled sentence embedding
> -----------------------------------------------------------------------------------------
>
>                 Key: OPENNLP-1977
>                 URL: https://issues.apache.org/jira/browse/OPENNLP-1977
>             Project: OpenNLP
>          Issue Type: Bug
>            Reporter: Richard Zowalla
>            Priority: Major
>             Fix For: 3.0.0
>
>
> {{SentenceVectorsDL.getVectors}} returns {{v\[0\]\[0\]}}, the raw last hidden state of the first \(\[CLS\]\) token. The class targets sentence\-transformers models converted to ONNX, and {{SentenceVectorsDLEval}} uses all\-MiniLM\-L6\-v2, whose {{1\_Pooling/config.json}} sets {{pooling\_mode\_mean\_tokens: true}} and {{pooling\_mode\_cls\_token: false}}, followed by a Normalize module. So the returned vector is not the embedding the model was trained to produce, and similarity scores differ from sentence\-transformers.
> Suggested fix: read the pooling mode and normalization from the model's sentence\-transformers metadata \({{1\_Pooling/config.json}}, {{modules.json}}\) when present, otherwise require it as an option; mean pooling must exclude padding positions. Since the 3.0.0 release note already states that vectors change, the default should be settled before 3.0.0.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)