Curie 7B V1
C
Curie 7B V1
Developed by szymonrucinski
Curie-7B-v1 is a model fine-tuned from English large language models (LLMs) for Polish text generation, excelling in Polish text generation and various NLP tasks.
Downloads 26
Release Time : 1/11/2024
Model Overview
This model became one of the best decoder models for generating Polish text through Language-Adaptive Pre-Training (LAPT) on high-quality Polish datasets and fine-tuning on the KLEJ benchmark.
Model Features
Efficient Language-Adaptive Pre-Training
Uses only 2-3% of the typically required data volume while achieving performance close to the best baseline models in multiple Polish NLP tasks.
High-Quality Polish Text Generation
Achieves a perplexity of 3.02, making it the best decoder model for generating Polish text.
Versatile Applications
Can be adapted as a classifier, regressor, or AI assistant, suitable for various Polish NLP tasks.
Model Capabilities
Polish text generation
Natural language processing tasks
Text classification
Sentiment analysis
Named entity recognition
Use Cases
Natural Language Processing
Polish Text Generation
Generates high-quality Polish text for content creation, translation assistance, and similar scenarios.
Perplexity of 3.02, with high-quality generated text.
Sentiment Analysis
Analyzes sentiment in Polish text, suitable for social media monitoring, customer feedback analysis, etc.
Achieves 92.7% accuracy on the PolEmo2.0-IN task.
Named Entity Recognition
Identifies named entities in Polish text for information extraction, knowledge graph construction, etc.
Achieves 93.4% accuracy on the NKJP-NER task.
Featured Recommended AI Models