Gliclass Base V1.0 Lw

🚀 ⭐ GLiClass: Generalist and Lightweight Model for Sequence Classification

GLiClass is an efficient zero - shot classifier inspired by the GLiNER work. It offers comparable performance to a cross - encoder while being more compute - efficient as classification is achieved in a single forward path. It can be applied to topic classification, sentiment analysis, and serve as a reranker in RAG pipelines. The model is trained on synthetic data and is suitable for commercial applications. This version uses layer - wise feature selection for better language understanding.

✨ Features

• Efficient Zero - Shot Classification: Performs classification in a single forward pass, offering high computational efficiency.
• Versatile Applications: Can be used for topic classification, sentiment analysis, and as a reranker in RAG pipelines.
• Commercial - Ready: Trained on synthetic data, making it suitable for commercial use.
• Advanced Feature Selection: Utilizes layer - wise feature selection for better language understanding.

📦 Installation

First of all, you need to install GLiClass library:

pip install gliclass

💻 Usage Examples

Basic Usage

from gliclass import GLiClassModel, ZeroShotClassificationPipeline
from transformers import AutoTokenizer

model = GLiClassModel.from_pretrained("knowledgator/gliclass-base-v1.0-lw")
tokenizer = AutoTokenizer.from_pretrained("knowledgator/gliclass-base-v1.0-lw")

pipeline = ZeroShotClassificationPipeline(model, tokenizer, classification_type='multi-label', device='cuda:0')

text = "One day I will see the world!"
labels = ["travel", "dreams", "sport", "science", "politics"]
results = pipeline(text, labels, threshold=0.5)[0] #because we have one text

for result in results:
 print(result["label"], "=>", result["score"])

📚 Documentation

Benchmarks

Below, you can see the F1 score on several text classification datasets. All tested models were not fine - tuned on those datasets and were tested in a zero - shot setting.

Below you can find a comparison with other GLiClass models:

Here you can see how the performance of the model grows providing more examples:

📄 License

This project is licensed under the Apache - 2.0 license.