Jasmine 350M
J
Jasmine 350M
Developed by UBC-NLP
JASMINE is a series of Arabic GPT models designed for few-shot learning, with parameters ranging from 300 million to 6.7 billion, pretrained on 235GB of text data.
Downloads 81
Release Time : 7/14/2022
Model Overview
JASMINE is an Arabic large language model based on the autoregressive Transformer architecture, specializing in few-shot learning tasks for Arabic text generation and comprehension.
Model Features
Arabic Language Optimization
Specially optimized for Arabic text, supporting high-quality generation and comprehension of Arabic.
Few-shot Learning
Designed for few-shot learning scenarios, excelling with limited training data.
Large-scale Pretraining
Pretrained on 235GB of Arabic text, equipped with rich linguistic knowledge.
Model Capabilities
Arabic text generation
Arabic text comprehension
Few-shot learning
Use Cases
Education
Arabic Learning Assistance
Helps learners understand and generate Arabic text
Content Creation
Arabic Poetry Generation
Generates Arabic poetry with philosophical and cultural characteristics
Example poetry in the style of Antara ibn Shaddad
Featured Recommended AI Models