German BERT large paraphrase cosine

This is a sentence-transformers model. It maps sentences & paragraphs (text) into a 1024 dimensional dense vector space. The model is intended to be used together with SetFit to improve German few-shot text classification. It has a sibling model called deutsche-telekom/gbert-large-paraphrase-euclidean.

This model is based on deepset/gbert-large. Many thanks to deepset!

Loss Function
We have used MultipleNegativesRankingLoss with cosine similarity as the loss function.

Training Data
The model is trained on a carefully filtered dataset of deutsche-telekom/ger-backtrans-paraphrase. We deleted the following pairs of sentences:

• min_char_len less than 15
• jaccard_similarity greater than 0.3
• de_token_count greater than 30
• en_de_token_count greater than 30
• cos_sim less than 0.85

Hyperparameters

• learning_rate: 8.345726930229726e-06
• num_epochs: 7
• train_batch_size: 57
• num_gpu: ???

Evaluation Results

We use the NLU Few-shot Benchmark - English and German dataset to evaluate this model in a German few-shot scenario.

Qualitative results

• multilingual sentence embeddings provide the worst results
• Electra models also deliver poor results
• German BERT base size model (deepset/gbert-base) provides good results
• German BERT large size model (deepset/gbert-large) provides very good results
• our fine-tuned models (this model and deutsche-telekom/gbert-large-paraphrase-euclidean) provide best results

Licensing

Copyright (c) 2023 Philip May, Deutsche Telekom AG
Copyright (c) 2022 deepset GmbH

Licensed under the MIT License (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License by reviewing the file LICENSE in the repository.