T

TURNA

Developed by boun-tabi-LMG
TURNA is a Turkish language model based on the UL2 framework, suitable for comprehension and generation tasks, demonstrating excellent performance in multiple Turkish language tasks.
Downloads 187
Release Time : 1/21/2024

Model Overview

TURNA is a Transformer-based encoder-decoder model specifically designed for Turkish. It outperforms multilingual models in various generation and comprehension tasks and competes with monolingual Turkish models.

Model Features

Turkish Language Optimization
Specifically designed and optimized for Turkish, excelling in Turkish language tasks
Multi-task Capability
Supports both generation and comprehension tasks, including paraphrasing, summarization, named entity recognition, etc.
Large-scale Pretraining
Pretrained on 10GB of Turkish text, exposed to 42.7 billion tokens

Model Capabilities

Text generation
Text comprehension
Paraphrase generation
Summarization
Named entity recognition
Part-of-speech tagging
Semantic text similarity analysis
Natural language inference
Text classification

Use Cases

Natural Language Processing
News Headline Generation
Automatically generates headlines based on news content
Performs well on TRNews and MLSUM datasets
Text Paraphrasing
Generates sentences with the same meaning but different expressions from the original text
Performs well on TAT and OST datasets
Text Analysis
Sentiment Analysis
Analyzes the sentiment tendency of product reviews and tweets
Performs well on Turkish product reviews and TTC4900 dataset
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase