M

M BERT Distil 40

Developed by M-CLIP
A model based on distilbert-base-multilingual, fine-tuned to align the embedding space for 40 languages, matching the embedding space of CLIP text encoder.
Downloads 46
Release Time : 3/2/2022

Model Overview

This is a multilingual text encoding model, fine-tuned to align with the embedding space of CLIP text encoder, supporting 40 languages.

Model Features

Multilingual Support
Supports text embedding for 40 languages, covering a wide range of linguistic diversity.
CLIP-Compatible
Fine-tuned to align with the embedding space of CLIP text encoder, enabling seamless integration with CLIP visual encoder.
Based on DistilBERT
Built upon the distilbert-base-multilingual-cased model, offering efficient performance.

Model Capabilities

Multilingual text embedding
Integration with CLIP visual encoder
Text processing for 40 languages

Use Cases

Multilingual Applications
Multilingual Image Captioning
Works with CLIP visual encoder to generate image captions in multiple languages.
Performs well in languages such as French, German, Spanish, Russian, Swedish, and Greek.
Cross-Language Search
Used for cross-language text and image search.
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase