German Semantic

🚀 German STS

A model for semantic sentence similarity in German, leveraging sentence-transformers technology.

🚀 Quick Start

✨ Features

• This model is tagged with semantic, sentence-transformers, and sentence-similarity, indicating its capabilities in semantic analysis and sentence similarity calculation.
• It is trained on the STS dataset, which helps it achieve good performance in German semantic tasks.

📦 Installation

To use this model, you need to install the sentence-transformers library. You can install it using the following command:

!pip install -U sentence-transformers

💻 Usage Examples

Basic Usage

The following is an example of using the model to calculate the cosine similarity between sentence pairs:

from sentence_transformers import SentenceTransformer
import torch
from sentence_transformers import util

model = SentenceTransformer('..model_path..')
sentences1 = ['Die Katze sitzt draußen',
             "Ein Mann spielt Gitarre",
             'Der neue Film ist großartig']

sentences2 = ['Der Hund spielt im Garten',
              "Eine Frau sieht fern",
              'Der neue Film ist so toll']
embeddings1 = model.encode(sentences1, convert_to_tensor=True)
embeddings2 = model.encode(sentences2, convert_to_tensor=True)
cosine_scores = util.pytorch_cos_sim(embeddings1, embeddings2)

for i in range(len(sentences1)):
    for j in range(len(sentences2)):
        print(cosine_scores[i][j])

# The output will be similar to the following:
# Die Katze sitzt draußen  Der Hund spielt im Garten  Score: 0.1259
# Die Katze sitzt draußen  Eine Frau sieht fern Score: 0.0567
# Die Katze sitzt draußen  Der neue Film ist so toll Score: 0.0557
# Ein Mann spielt Gitarre  Der Hund spielt im Garten  Score: 0.1031
# Ein Mann spielt Gitarre  Eine Frau sieht fern Score: 0.0098
# Ein Mann spielt Gitarre  Der neue Film ist so toll  Score: 0.0828
# Der neue Film ist großartig Der Hund spielt im Garten  Score: 0.1008
# Der neue Film ist großartig Eine Frau sieht fern  Score: 0.0674

📚 Documentation

Performance on STS Datasets

• STS dev (German): 87.9%
• STS test (German): 84.3%

🔧 Technical Details

This model is based on the sentence-transformers library, which is a powerful tool for semantic analysis and sentence similarity calculation. It uses a pre - trained model to encode sentences into vectors and then calculates the cosine similarity between these vectors to measure the semantic similarity between sentences.

📄 Information Table