SentenceTransformer based on NbAiLab/nb-bert-base

This is a sentence-transformers model finetuned from NbAiLab/nb-bert-base. It is the second version of the existing NbAiLab/nb-sbert-base model, providing a larger max sequence length for inputs.

The model maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. The easiest way is to simply measure the cosine distance between two sentences. Sentences that are close to each other in meaning, will have a small cosine distance and a similarity close to 1. The model is trained in such a way that similar sentences in different languages should also be close to each other. Ideally, an English-Norwegian sentence pair should have high similarity.

Model Details

Model Description

Model Type: Sentence Transformer
Base model: NbAiLab/nb-bert-base
Maximum Sequence Length: 512 tokens
Output Dimensionality: 768 dimensions
Similarity Function: Cosine Similarity
Training Dataset: Subset of NbAiLab/mnli-norwegian
Language: Norwegian and English
License: Apache 2.0

EU AI Act

This release is a non-generative encoder model whose outputs are vectors/scores rather than language or media. Its intended functionality is limited to representation, retrieval, ranking, or classification support. On that basis, the release is preliminarily assessed as not falling within the provider obligations for GPAI models under the EU AI Act definitions, subject to legal confirmation if capability scope or marketed generality changes. For more information, see the Model Documentation Form here.

Model Sources

Documentation: Sentence Transformers Documentation
Repository: Sentence Transformers on GitHub
Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': False, 'architecture': 'BertModel'})
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
)

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load this model and run inference.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("NbAiLab/nb-sbert-v2-base")
# Run inference
sentences = [
    "This is a Norwegian boy", 
    "Dette er en norsk gutt"
    ]

embeddings = model.encode(sentences)
print(embeddings.shape)
# (2, 768)

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 0.8287],
#         [0.8287, 1.0000]])

Direct Usage (Transformers)

Without sentence-transformers, you can still use the model. First, you pass in your input through the transformer model, then you have to apply the right pooling-operation on top of the contextualized word embeddings.

Click to see the direct usage in Transformers

import torch

from sklearn.metrics.pairwise import cosine_similarity
from transformers import AutoTokenizer, AutoModel

#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
    token_embeddings = model_output[0] #First element of model_output contains all token embeddings
    input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
    return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)


# Sentences we want sentence embeddings for
sentences = ["This is a Norwegian boy", "Dette er en norsk gutt"]

# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('NbAiLab/nb-sbert-v2-base')
model = AutoModel.from_pretrained('NbAiLab/nb-sbert-v2-base')

# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')

# Compute token embeddings
with torch.no_grad():
    model_output = model(**encoded_input)

# Perform pooling. In this case, mean pooling.
embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
print(embeddings.shape)
# torch.Size([2, 768])

similarity = cosine_similarity(embeddings[0].reshape(1, -1), embeddings[1].reshape(1, -1))
print(similarity)
# This should give 0.8287 in the example above.

Evaluation

Metrics

Semantic Similarity

Dataset: STS Benchmark
Evaluated with EmbeddingSimilarityEvaluator

Metric	nb-sbert-base	nb-sbert-v2-base
pearson_cosine	0.8275	0.8478
spearman_cosine	0.8245	0.8495

MTEB (Scandinavian)

Metric	nb-sbert-base	nb-sbert-v2-base
Mean (Task)	0.5190	0.5496
Mean (TaskType)	0.5394	0.5690

Bitext Mining	0.7228	0.7275
Classification	0.5708	0.5841
Clustering	0.3798	0.4105
Retrieval	0.4840	0.5540

Training Details

Training Dataset

Subset of NbAiLab/mnli-norwegian

Size: 527,098 training samples
Columns: anchor, positive, and negative

Approximate statistics based on the first 1000 samples:

	anchor	positive	negative
type	string	string	string
details	min: 5 tokens mean: 20.91 tokens max: 130 tokens	min: 5 tokens mean: 20.91 tokens max: 130 tokens	min: 5 tokens mean: 14.14 tokens max: 39 tokens

Samples:

anchor	positive	negative
`Det som følger er mindre en glid nedover en glatt skråning enn et profesjonelt skred som resulterer i enten en oppsigelse eller en smal flukt til neste drømmejobb, der, selvfølgelig, syklusen gjentas igjen.`	`Syklusen gjentar seg ved neste jobb.`	`Syklusen gjentar seg sjelden ved neste jobb.`
`Syklusen gjentar seg ved neste jobb.`	`Det som følger er mindre en glid nedover en glatt skråning enn et profesjonelt skred som resulterer i enten en oppsigelse eller en smal flukt til neste drømmejobb, der, selvfølgelig, syklusen gjentas igjen.`	`Syklusen gjentar seg sjelden ved neste jobb.`
`The public areas are spectacular, the rooms a bit less so, but a long-awaited renovation was carried out in 1998.`	`The rooms are nice, but the public area is in a league of it's own.`	`The public area was fine, but the rooms were really something else.`
`Ah, but he had no opportunity.`	`Han hadde ikke sjansen til å gjøre noe.`	`Han hadde mange muligheter.`

Loss: MultipleNegativesRankingLoss with these parameters:

{
    "scale": 20.0,
    "similarity_fct": "cos_sim",
    "gather_across_devices": false
}

Training Hyperparameters

Non-Default Hyperparameters

per_device_train_batch_size: 128
num_train_epochs: 1
learning_rate: 2e-05
warmup_steps: 412.0
bf16: True
eval_strategy: steps
per_device_eval_batch_size: 128
batch_sampler: no_duplicates

All Hyperparameters

Click to expand

per_device_train_batch_size: 128
num_train_epochs: 1
max_steps: -1
learning_rate: 2e-05
lr_scheduler_type: linear
lr_scheduler_kwargs: None
warmup_steps: 412.0
optim: adamw_torch_fused
optim_args: None
weight_decay: 0.0
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1e-08
optim_target_modules: None
gradient_accumulation_steps: 1
average_tokens_across_devices: True
max_grad_norm: 1.0
label_smoothing_factor: 0.0
bf16: True
fp16: False
bf16_full_eval: False
fp16_full_eval: False
tf32: None
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
torch_compile: False
torch_compile_backend: None
torch_compile_mode: None
use_liger_kernel: False
liger_kernel_config: None
use_cache: False
neftune_noise_alpha: None
torch_empty_cache_steps: None
auto_find_batch_size: False
log_on_each_node: True
logging_nan_inf_filter: True
include_num_input_tokens_seen: no
log_level: passive
log_level_replica: warning
disable_tqdm: False
project: huggingface
trackio_space_id: trackio
eval_strategy: steps
per_device_eval_batch_size: 128
prediction_loss_only: True
eval_on_start: False
eval_do_concat_batches: True
eval_use_gather_object: False
eval_accumulation_steps: None
include_for_metrics: []
batch_eval_metrics: False
save_only_model: False
save_on_each_node: False
enable_jit_checkpoint: False
push_to_hub: False
hub_private_repo: None
hub_model_id: None
hub_strategy: every_save
hub_always_push: False
hub_revision: None
load_best_model_at_end: False
ignore_data_skip: False
restore_callback_states_from_checkpoint: False
full_determinism: False
seed: 42
data_seed: None
use_cpu: False
accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
parallelism_config: None
dataloader_drop_last: False
dataloader_num_workers: 0
dataloader_pin_memory: True
dataloader_persistent_workers: False
dataloader_prefetch_factor: None
remove_unused_columns: True
label_names: None
train_sampling_strategy: random
length_column_name: length
ddp_find_unused_parameters: None
ddp_bucket_cap_mb: None
ddp_broadcast_buffers: False
ddp_backend: None
ddp_timeout: 1800
fsdp: []
fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
deepspeed: None
debug: []
skip_memory_metrics: True
do_predict: False
resume_from_checkpoint: None
warmup_ratio: None
local_rank: -1
prompts: None
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Click to see expand

| Epoch | Step | Training Loss | sts-dev_spearman_cosine | |:------:|:----:|:-------------:|:-----------------------:| | 0.0243 | 100 | 1.8923 | - | | 0.0486 | 200 | 0.8525 | - | | 0.0729 | 300 | 0.6760 | - | | 0.0971 | 400 | 0.5891 | - | | 0.1000 | 412 | - | 0.8397 | | 0.1214 | 500 | 0.5567 | - | | 0.1457 | 600 | 0.5245 | - | | 0.1700 | 700 | 0.5024 | - | | 0.1943 | 800 | 0.4654 | - | | 0.2001 | 824 | - | 0.8390 | | 0.2186 | 900 | 0.4796 | - | | 0.2428 | 1000 | 0.4528 | - | | 0.2671 | 1100 | 0.4577 | - | | 0.2914 | 1200 | 0.4443 | - | | 0.3001 | 1236 | - | 0.8455 | | 0.3157 | 1300 | 0.4201 | - | | 0.3400 | 1400 | 0.4010 | - | | 0.3643 | 1500 | 0.4063 | - | | 0.3885 | 1600 | 0.3955 | - | | 0.4002 | 1648 | - | 0.8446 | | 0.4128 | 1700 | 0.3798 | - | | 0.4371 | 1800 | 0.3772 | - | | 0.4614 | 1900 | 0.3933 | - | | 0.4857 | 2000 | 0.3793 | - | | 0.5002 | 2060 | - | 0.8499 | | 0.5100 | 2100 | 0.3862 | - | | 0.5342 | 2200 | 0.3730 | - | | 0.5585 | 2300 | 0.3463 | - | | 0.5828 | 2400 | 0.3556 | - | | 0.6003 | 2472 | - | 0.8503 | | 0.6071 | 2500 | 0.3614 | - | | 0.6314 | 2600 | 0.3479 | - | | 0.6557 | 2700 | 0.3508 | - | | 0.6799 | 2800 | 0.3463 | - | | 0.7003 | 2884 | - | 0.8471 | | 0.7042 | 2900 | 0.3453 | - | | 0.7285 | 3000 | 0.3327 | - | | 0.7528 | 3100 | 0.3269 | - | | 0.7771 | 3200 | 0.3333 | - | | 0.8004 | 3296 | - | 0.8493 | | 0.8014 | 3300 | 0.3370 | - | | 0.8256 | 3400 | 0.3254 | - | | 0.8499 | 3500 | 0.3348 | - | | 0.8742 | 3600 | 0.3213 | - | | 0.8985 | 3700 | 0.3376 | - | | 0.9004 | 3708 | - | 0.8495 | | 0.9228 | 3800 | 0.3362 | - | | 0.9471 | 3900 | 0.3246 | - | | 0.9713 | 4000 | 0.3215 | - | | 0.9956 | 4100 | 0.3143 | - |

Framework Versions

Python: 3.14.3
Sentence Transformers: 5.2.3
Transformers: 5.3.0
PyTorch: 2.10.0+cu130
Accelerate: 1.13.0
Datasets: 4.6.1
Tokenizers: 0.22.2

Citation

BibTeX

Sentence Transformers

@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",
}

MultipleNegativesRankingLoss

@misc{henderson2017efficient,
    title={Efficient Natural Language Response Suggestion for Smart Reply},
    author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
    year={2017},
    eprint={1705.00652},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}

NbAiLab/nb-bert-base

@inproceedings{kummervold-etal-2021-operationalizing,
  title     = {Operationalizing a National Digital Library: The Case for a {N}orwegian Transformer Model},
  author    = {Kummervold, Per E  and
               De la Rosa, Javier  and
               Wetjen, Freddy  and
               Brygfjeld, Svein Arne},
  booktitle = {Proceedings of the 23rd Nordic Conference on Computational Linguistics (NoDaLiDa)},
  year      = {2021},
  address   = {Reykjavik, Iceland (Online)},
  publisher = {Linköping University Electronic Press, Sweden},
  url       = {https://huggingface.co/papers/2104.09617},
  pages     = {20--29},
  abstract  = {In this work, we show the process of building a large-scale training set from digital and digitized collections at a national library. The resulting Bidirectional Encoder Representations from Transformers (BERT)-based language model for Norwegian outperforms multilingual BERT (mBERT) models in several token and sequence classification tasks for both Norwegian Bokmål and Norwegian Nynorsk. Our model also improves the mBERT performance for other languages present in the corpus such as English, Swedish, and Danish. For languages not included in the corpus, the weights degrade moderately while keeping strong multilingual properties. Therefore, we show that building high-quality models within a memory institution using somewhat noisy optical character recognition (OCR) content is feasible, and we hope to pave the way for other memory institutions to follow.},
}

Citing & Authors

The model was trained by Victoria Handford and Lucas Georges Gabriel Charpentier. The documentation was initially autogenerated by the SentenceTransformers library then revised by Victoria Handford, Lucas Georges Gabriel Charpentier, and Javier de la Rosa.