Downloads · 30 days
42
8% of all-time downloads
vincolle/reranker-bert-italian-uncased-mmarco-mnrl
reranker-bert-italian-uncased-mmarco-mnrl is a text ranking model from vincolle. Use it for the text ranking task on the model card, and read the license before you ship it in a product. It is set up for sentence-transformers. The card lists the license as apache-2.0.
<p align="center" <img src="images/logo.png" alt="Italian ModernBERT Reranker" width="600"/ </p
Downloads · 30 days
42
8% of all-time downloads
All-time downloads
511
Public
Parameters
135M
556 MB on disk
Likes
2
Public
Click a slice to open those files.
.safetensors542 MB · 98%
From the Hugging Face model README
Italian-language information retrieval has very few dedicated reranker models — most pipelines either fall back to multilingual cross-encoders (diluted by 100 languages) or skip reranking entirely. This model is trained natively in Italian, on Italian-specific hard negatives, to fill that gap.
Evaluated with CrossEncoderRerankingEvaluator on a held-out 2,500-sample dev split of the training distribution:
| Metric | Score | Δ over base retriever |
|---|---|---|
| MAP | 0.9105 | +0.0703 |
| MRR@10 | 0.9102 | +0.0712 |
| NDCG@10 | 0.9235 | +0.0576 |
"Base retriever" refers to nickprock/multi-sentence-BERTino — a strong, Italian-specialized bi-encoder — used to mine hard negatives during training. The gain reported here is therefore a realistic estimate of what this reranker adds on top of a competitive existing Italian retriever, not a comparison against a weak baseline.
The model was trained on a high-quality Italian passage ranking mixture combining samples from mMARCO-it and vincolle/gooaq-italian-80k, an Italian translation of GooAQ that I produced myself (translated with DeepSeek). To my knowledge this is the first public Italian version of GooAQ — it does not exist as a native dataset in any other language besides English.
Hard negatives were mined using nickprock/multi-sentence-BERTino, a native Italian bi-encoder trained with Matryoshka Representation Learning. Using an Italian-specific embedder for mining (rather than a generic multilingual one) was a deliberate choice — early experiments with multilingual embedders produced negatives that were either too easy (trivially distinguishable) or, in the worst case, contained answers just as valid as the labeled positive. Switching to an Italian-native retriever for mining produced negatives that were genuinely hard without being false negatives, which is what made margin-based filtering during mining meaningful at all.
pip install -U sentence-transformers
from sentence_transformers import CrossEncoder
model = CrossEncoder("vincolle/reranker-bert-italian-uncased-mmarco-mnrl")
# Score query-document pairs directly
pairs = [
("Quante calorie si bruciano correndo 5 miglia?",
"Correndo 5 miglia si brucia una grande quantità di calorie. In generale, la maggior parte delle persone che corre 5 miglia brucia circa 500 calorie."),
("Quante calorie si bruciano correndo 5 miglia?",
"Le pinzette sono consentite nel bagaglio a mano, così come i rasoi elettrici."),
]
scores = model.predict(pairs)
print(scores)
# [0.99..., 0.05...]
# Or rerank a list of candidate documents for a single query
ranks = model.rank(
"Quante calorie si bruciano correndo 5 miglia?",
[
"Correndo 5 miglia si brucia una grande quantità di calorie...",
"Le pinzette sono consentite nel bagaglio a mano...",
"Guidare con ammortizzatori difettosi può essere estremamente pericoloso...",
],
)
This reranker is designed as the second stage of a retrieve-then-rerank pipeline:
nickprock/multi-sentence-BERTino)from sentence_transformers import SentenceTransformer, CrossEncoder
retriever = SentenceTransformer("nickprock/multi-sentence-BERTino")
reranker = CrossEncoder("vincolle/reranker-bert-italian-uncased-mmarco-mnrl")
query = "Quante calorie si bruciano correndo 5 miglia?"
candidates = retriever.similarity(...) # your top-k retrieval step
reranked = reranker.rank(query, candidates)
| Parameter | Configuration |
|---|---|
| Loss Function | CachedMultipleNegativesRankingLoss |
| Optimization | AdamW (adamw_torch_fused), mixed precision (fp16) |
| Batch Size | 128 (per device) |
| Learning Rate | 2e-5 (linear schedule, 0.1 warmup ratio) |
| Epochs | 1 |
| Random Seed | 42 |
| Step | Training Loss | Italian-Dev NDCG@10 |
|---|---|---|
| 1 | 2.4481 | – |
| 100 | 1.3025 | 0.8877 |
| 200 | 0.4025 | 0.9110 |
| 300 | 0.3212 | 0.9165 |
| 400 | 0.2890 | 0.9197 |
| 500 | 0.2640 | 0.9220 |
| 586 (final) | – | 0.9235 |
If you use this model, please cite the base model and the Sentence Transformers library:
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "https://arxiv.org/abs/1908.10084",
}