# Contrastive Learning

This is a pre-trained backdoor-injected model for studying backdoor sample detection in contrastive language-image pretraining.

A SigLIP (Sigmoid Loss for Language-Image Pre-training) model trained on the multilingual WebLI dataset, supporting multilingual image classification tasks.

A Vision Transformer model for detecting deepfake images, achieving high-precision forgery detection through contrastive learning and global-local similarity analysis.

A Japanese version of the CLIP model that maps Japanese text and images into the same embedding space, suitable for zero-shot image classification, text-image retrieval, and other tasks.

CLIPA-v2 model, an efficient contrastive vision-language model focused on zero-shot image classification tasks.

WRAPresentations is an advanced sentence transformer model specifically designed for Twitter argument mining, capable of mapping tweets into four categories: 'Reason', 'Statement', 'Notification', and 'Nonsense'.

Sentence embedding model based on MPNet architecture, mapping text to a 384-dimensional vector space, suitable for semantic search and sentence similarity tasks

Sentence embedding model based on MPNet architecture, mapping text to a 768-dimensional vector space, suitable for semantic search and text similarity tasks

ClipMD is a medical image-text matching model developed based on OpenAI's CLIP model, employing a sliding window text encoder specifically designed for medical image classification tasks.

miCSE is a few-shot sentence embedding model specifically designed for sentence similarity calculation, achieving efficient sample learning through attention pattern alignment and regularization of self-attention distributions.

X-CLIP is a minimalist extension of CLIP for general video-language understanding, trained contrastively on (video, text) pairs, suitable for zero-shot, few-shot, or fully supervised video classification as well as video-text retrieval tasks.

X-CLIP is an extended version of CLIP for general video-language understanding, supporting video classification and video-text retrieval tasks.

X-CLIP is an extended version of CLIP for general video-language understanding, trained via contrastive learning on video-text pairs, supporting zero-shot, few-shot, and fully supervised video classification tasks.

Sentence embedding model based on MPNet architecture, mapping text to a 768-dimensional vector space, suitable for semantic search and text similarity tasks

This is a sentence embedding model based on sentence-transformers, capable of mapping text to a 768-dimensional vector space, suitable for semantic search and clustering tasks.

This is a sentence embedding model based on sentence-transformers, capable of mapping text to a 384-dimensional vector space, suitable for semantic search and clustering tasks.

This is a sentence embedding model based on the MPNet architecture, capable of mapping sentences and paragraphs into a 768-dimensional dense vector space, suitable for tasks like semantic search and clustering.

A contrastive learning model trained using parallel language pairs, mapping texts to the same vector space across different languages

Unsupervised contrastive learning model based on BERT architecture, improving sentence embedding quality through a simple yet effective contrastive learning framework

A simplified Chinese sentence embedding encoding model based on simple contrastive learning, using the Chinese RoBERTa WWM extended version as the pre-trained model.

SimeCSE_Vietnamese is a Vietnamese sentence embedding model based on SimCSE, using PhoBERT as the pretrained language model, suitable for both unlabeled and labeled data.

Supervised SimCSE model based on RoBERTa-large for sentence embedding and feature extraction tasks.

A fine-tuned Facebook wav2vec2 model for the speech-to-text module of The Sound of AI Open Source Research Group

A fine-tuned version of the Contriever pre-trained model, optimized for dense information retrieval tasks and trained using contrastive learning methods

CODER is a knowledge-enhanced cross-language medical term embedding model, focusing on the task of medical terminology standardization.