S

Sea Lion 3b

Developed by aisingapore
SEA-LION is a series of multilingual large language models optimized for Southeast Asia, supporting 11 Southeast Asian languages, built on the MPT architecture.
Downloads 770
Release Time : 10/24/2023

Model Overview

SEA-LION-v1-3B is a 3-billion-parameter large language model specifically pre-trained and instruction-tuned for Southeast Asian languages, supporting English, Chinese, and 9 local Southeast Asian languages.

Model Features

Southeast Asian language optimization
Specifically trained for 11 Southeast Asian languages, including low-resource languages such as Lao and Khmer.
Large vocabulary support
Utilizes the 256K-vocabulary SEABPETokenizer for optimized tokenization of Southeast Asian languages.
Large-scale training data
Trained on 980 billion tokens of multilingual data, covering various languages and technical domains.
Efficient architecture
Optimized based on the MPT architecture with a 32-layer decoder structure, suitable for Southeast Asian language processing.

Model Capabilities

Multilingual text generation
Cross-language understanding
Code generation (Python/Javascript, etc.)
Technical document processing

Use Cases

Multilingual applications
Southeast Asian customer service chatbot
Supports automated customer service systems in multiple Southeast Asian languages.
Capable of handling customer inquiries in 11 languages.
Multilingual content generation
Generates localized marketing content for the Southeast Asian market.
Supports ad copy and social media content creation in multiple languages.
Technical applications
Code assistance
Supports code generation and completion for languages like Python and Javascript.
Helps developers improve coding efficiency.
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase