PolDense-68M / README.md
sdadas's picture
Update README.md
1ae4537 verified
|
Raw
History Blame Contribute Delete
87.2 kB
---
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
language: pl
license: gemma
widget:
- source_sentence: "zapytanie: Jak dożyć 100 lat?"
sentences:
- "Trzeba zdrowo się odżywiać i uprawiać sport."
- "Trzeba pić alkohol, imprezować i jeździć szybkimi autami."
- "Gdy trwała kampania politycy zapewniali, że rozprawią się z zakazem niedzielnego handlu."
---
<h1 align="center">PolDense-68M</h1>
PolDense is a family of Polish embedding models optimized for dense retrieval. The models are built on top of the [ettin-encoders](https://huggingface.co/collections/jhu-clsp/encoders-vs-decoders-the-ettin-suite), which use the ModernBERT architecture, and are designed to provide high-quality text representations for Polish search, retrieval, and RAG applications. The family consists of six models, ranging from 17M to 1B parameters. This makes it possible to choose between lightweight, low-latency deployments and larger models aimed at maximum retrieval quality.
## Training
PolDense models were trained in a three-stage pipeline with two knowledge distillation stages and one fine-tuning stage. The main teacher embedding model was [BGE-Multilingual-Gemma2](https://huggingface.co/BAAI/bge-multilingual-gemma2), while the final retrieval fine-tuning stage used relevance scores produced by [BGE-Reranker-v2.5-Gemma2-Lightweight](https://huggingface.co/BAAI/bge-reranker-v2.5-gemma2-lightweight).
1. The first stage adapted the original English [ettin-encoder](https://huggingface.co/collections/jhu-clsp/encoders-vs-decoders-the-ettin-suite) models to Polish using multilingual knowledge distillation. This stage follows the approach introduced in [Making Monolingual Sentence Embeddings Multilingual using Knowledge Distillation](https://arxiv.org/abs/2004.09813). The core idea is to use a strong teacher model to define a shared semantic embedding space. For each parallel sentence pair, the teacher encodes the source-language sentence, and the student is trained to produce a similar embedding for the translated sentence. In practice, this teaches the student to map Polish and English texts with the same meaning to nearby locations in the vector space, while preserving the retrieval-oriented structure learned by the teacher. In this stage, [BGE-Multilingual-Gemma2](https://huggingface.co/BAAI/bge-multilingual-gemma2) was used as the teacher model. Training was performed on a Polish-English parallel corpus containing approximately 20 million text pairs. Each model was trained for 5 epochs with a batch size of 64.
2. The second stage used a more advanced distillation objective inspired by [Jasper and Stella: distillation of SOTA embedding models](https://arxiv.org/abs/2412.19048). The goal of this stage was to align the student models more closely with the teacher and extract as much retrieval quality as possible from the distillation process. Unlike the first stage, this stage was performed only on Polish texts. The training objective combined three complementary losses: a) a cosine alignment loss; b) a pairwise similarity loss; c) a triplet-style ranking loss. Together, these losses train the student not only to imitate individual teacher vectors, but also to reproduce the teacher's local geometry and ranking behavior. This is particularly important for retrieval, where relative similarity between queries and passages matters more than isolated embedding values. The teacher model in this stage was again [BGE-Multilingual-Gemma2](https://huggingface.co/BAAI/bge-multilingual-gemma2). Models from 17M to 150M parameters were trained on approximately 70 million Polish texts, while the 400M and 1B models were trained on approximately 33 million Polish texts. The corpus included the 20 million Polish texts used in first stage, with additional texts taken from the Polish portion of the [FineTranslations](https://huggingface.co/datasets/HuggingFaceFW/finetranslations) collection. Each model was trained for 5 epochs with a batch size of 128.
3. The final stage fine-tuned the models directly for retrieval using contrastive learning. This stage used 13 retrieval training datasets containing over 4.5 million queries and more than 15 million passages. No ground-truth labels were used during this stage. Instead, the training data was constructed entirely from positive and negative examples selected with the [BGE-Reranker-v2.5-Gemma2-Lightweight](https://huggingface.co/BAAI/bge-reranker-v2.5-gemma2-lightweight) reranker. The models were trained for 10 epochs with a batch size of 1024.
The figure below presents an evaluation of PolDense models on the [PIRB benchmark](https://huggingface.co/spaces/sdadas/pirb), which consists of 41 Polish retrieval tasks. As shown, each training stage contributed to the model’s final performance:
<center>
<img src="data:image/png;base64,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" />
</center>
## Evaluation
The PolDense models were evaluated on the PIRB benchmark. The results show that they outperform previously available retrievers for Polish, with each model achieving the best performance in its size class. The smaller 32M and 68M variants reach results comparable to much larger multilingual models with several hundred million parameters, such as [Jina-Embeddings-V5](https://huggingface.co/jinaai/jina-embeddings-v5-text-small), [BGE-M3](https://huggingface.co/BAAI/bge-m3), and [Snowflake-Arctic-Embed-2.0](https://huggingface.co/Snowflake/snowflake-arctic-embed-l-v2.0). The 150M model performs close to [Qwen3-Embedding-8B](https://huggingface.co/Qwen/Qwen3-Embedding-8B), while the 400M model reaches a level similar to [BGE-Multilingual-Gemma2](BAAI/bge-multilingual-gemma2) and [Llama-Embed-Nemotron-8b](https://huggingface.co/nvidia/llama-embed-nemotron-8b). The largest 1B variant surpasses both of these models. The results are shown in the figure below.
<center>
<img src="data:image/png;base64,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" />
</center>
## Usage
The model utilizes prefixes for retrieval and semantic similarity tasks. For retrieval, queries should be prefixed with **"[query]: "**. For symmetric tasks such as semantic similarity, both texts should be prefixed with **"[sts]: "**.
We recommend using the model with sentence-transformers version 5.4.0 or newer and initializing it with the arguments `dtype="float16"` or `"bfloat16"`, and `attn_implementation="flash_attention_2"`, which enables optimal performance. You can use the model with sentence-transformers:
```python
from sentence_transformers import SentenceTransformer
from sentence_transformers.util import cos_sim
model = SentenceTransformer(
"OPI-PIB/PolDense-68M",
device="cuda",
model_kwargs={
"dtype": "bfloat16",
"attn_implementation": "flash_attention_2"
}
)
# Retrieval example
query_prefix = "[query]: "
queries = [query_prefix + "Jak dożyć 100 lat?"]
answers = [
"Trzeba zdrowo się odżywiać i uprawiać sport.",
"Trzeba pić alkohol, imprezować i jeździć szybkimi autami.",
"Gdy trwała kampania politycy zapewniali, że rozprawią się z zakazem niedzielnego handlu."
]
queries_emb = model.encode(queries, convert_to_tensor=True, show_progress_bar=False)
answers_emb = model.encode(answers, convert_to_tensor=True, show_progress_bar=False)
best_answer = cos_sim(queries_emb, answers_emb).argmax().item()
print(answers[best_answer])
# Semantic similarity example
sim_prefix = "[sts]: "
sentences = [
sim_prefix + "Trzeba zdrowo się odżywiać i uprawiać sport.",
sim_prefix + "Warto jest prowadzić zdrowy tryb życia, uwzględniający aktywność fizyczną i dietę.",
sim_prefix + "One should eat healthy and engage in sports.",
sim_prefix + "Zakupy potwierdzasz PINem, który bezpiecznie ustalisz podczas aktywacji."
]
emb = model.encode(sentences, convert_to_tensor=True, show_progress_bar=False)
print(cos_sim(emb, emb))
```
Running with vLLM:
```sh
vllm serve OPI-PIB/PolDense-68M --dtype bfloat16 --runner pooling --convert embed --max-model-len 7999
```
## Acknowledgements
The research was supported by the project <strong>Large Language Models for the European Union (LLMs4EU)</strong>. This project is co-funded by the Digital Europe Programme under Grant Agreement 101198470.
The research was supported [in part] by project “<strong>Cloud Artificial Intelligence Service Engineering (CAISE)</strong> platform to create universal and smart services for various application areas”, No. KPOD.05.10-IW.10-0005/24, as part of the European IPCEI-CIS program, financed by NRRP (National Recovery and Resilience Plan) funds. Computations were carried out using the computers of <strong>Centre of Informatics Tricity Academic Supercomputer & Network at Gdansk University of Technology</strong>.