SAP
/

miCSE

Sentence Similarity

feature-extraction

text-embeddings-inference

Model card Files Files and versions

TJKlein commited on Nov 19, 2022

Commit

16eb4c5

·

1 Parent(s): d6917ea

Update README.md

Files changed (1) hide show

README.md +1 -1

README.md CHANGED Viewed

@@ -6,7 +6,7 @@ license: apache-2.0
 Language model of the pre-print arXiv paper titled: "_**miCSE**: Mutual Information Contrastive Learning for Low-shot Sentence Embeddings_"
 The **miCSE** language model is trained for sentence similarity computation. Training the model imposes alignment between the attention pattern of different views (embeddings of augmentations) during contrastive learning. Learning sentence embeddings with **miCSE** entails enforcing the syntactic consistency across augmented views for every single sentence, making contrastive self-supervised learning more sample efficient. This is achieved by regularizing the attention distribution. Regularizing the attention space enables learning representation in self-supervised fashion even when the training corpus is comparatively small.

 Language model of the pre-print arXiv paper titled: "_**miCSE**: Mutual Information Contrastive Learning for Low-shot Sentence Embeddings_"
+# Brief Model Description
 The **miCSE** language model is trained for sentence similarity computation. Training the model imposes alignment between the attention pattern of different views (embeddings of augmentations) during contrastive learning. Learning sentence embeddings with **miCSE** entails enforcing the syntactic consistency across augmented views for every single sentence, making contrastive self-supervised learning more sample efficient. This is achieved by regularizing the attention distribution. Regularizing the attention space enables learning representation in self-supervised fashion even when the training corpus is comparatively small.