# Multimodal Embedding

UniME is a general embedding learning model based on a multimodal large model, focusing on breaking down modal barriers to achieve cross-modal retrieval and embedding learning.

A multimodal embedding model based on Qwen2.5-Omni-7B, supporting unified embedding representations for cross-lingual text, images, audio, and video

Nomic Embed Multimodal 3B is a cutting-edge multimodal embedding model focused on visual document retrieval tasks, supporting unified text-image encoding, achieving an outstanding performance of 58.8 NDCG@5 in the Vidore-v2 test.

ColNomic Embed Multimodal 3B is a 3-billion-parameter multimodal embedding model specifically designed for visual document retrieval tasks, supporting unified encoding of multilingual text and images.

The first retriever specifically designed for financial time series forecasting, based on the Retrieval-Augmented Generation (RAG) framework

A fine-tuned version of BAAI/bge-m3 on Thai legal query data, supporting dense retrieval, lexical matching, and multi-vector interaction

LLaVE-7B is a 7-billion-parameter multimodal embedding model based on LLaVA-OneVision-7B, capable of embedding representations for text, images, multiple images, and videos.

LLaVE-2B is a 2-billion-parameter multimodal embedding model based on Aquila-VL-2B, featuring a 4K token context window and supporting embeddings for text, images, multiple images, and videos.

LLaVE is a multimodal embedding model based on the LLaVA-OneVision-0.5B model, with a parameter scale of 0.5B, capable of embedding text, images, multiple images, and videos.

Vision Transformer model based on SigLIP architecture, containing only the image encoder part, using original attention pooling mechanism

DSE-QWen2-2b-MRL-V1 is a dual-encoder model specifically designed for encoding document screenshots into dense vectors to facilitate document retrieval.

GGUF quantized version of the bge-m3 embedding model, suitable for efficient text embedding tasks

High-performance visual embedding model, sharing the same embedding space with nomic-embed-text-v1.5, supporting multimodal applications

High-performance vision embedding model, sharing the same embedding space with nomic-embed-text-v1, supporting multimodal applications

BGE-M3 is an embedding model that supports dense retrieval, lexical matching, and multi-vector interaction, converted to ONNX format for compatibility with frameworks like ONNX Runtime.

SigLIP is a vision-language pre-trained model suitable for zero-shot image classification tasks.

OpenAI's open-source CLIP model, based on Vision Transformer architecture, supporting cross-modal understanding of images and text

The base version of Chinese CLIP, using ViT-B/16 as the image encoder and RoBERTa-wwm-base as the text encoder, trained on a large-scale dataset of approximately 200 million Chinese image-text pairs.