Update README.md

b737706 verified 7 days ago

26.7 kB

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - dense
  - generated_from_trainer
  - dataset_size:527098
  - loss:MultipleNegativesRankingLoss
base_model: NbAiLab/nb-bert-large
datasets: NbAiLab/mnli-norwegian
license: apache-2.0
language:
  - 'no'
widget:
  - source_sentence: The man talked to a girl over the internet camera.
    sentences:
      - A group of elderly people pose around a dining table.
      - A teenager talks to a girl over a webcam.
      - There is no 'still' that is not relative to some other object.
  - source_sentence: A woman is writing something.
    sentences:
      - Two eagles are perched on a branch.
      - >-
        It refers to the maximum f-stop (which is defined as the ratio of focal
        length to effective aperture diameter).
      - A woman is chopping green onions.
  - source_sentence: The player shoots the winning points.
    sentences:
      - Minimum wage laws hurt the least skilled, least productive the most.
      - The basketball player is about to score points for his team.
      - Sheep are grazing in the field in front of a line of trees.
  - source_sentence: >-
      Stars form in star-formation regions, which itself develop from molecular
      clouds.
    sentences:
      - >-
        Although I believe Searle is mistaken, I don't think you have found the
        problem.
      - >-
        It may be possible for a solar system like ours to exist outside of a
        galaxy.
      - >-
        A blond-haired child performing on the trumpet in front of a house while
        his younger brother watches.
  - source_sentence: >-
      While Queen may refer to both Queen regent (sovereign) or Queen consort,
      the King has always been the sovereign.
    sentences:
      - At first, I thought this is a bit of a tricky question.
      - A man sitting on the floor in a room is strumming a guitar.
      - >-
        There is a very good reason not to refer to the Queen's spouse as "King"
        - because they aren't the King.
pipeline_tag: sentence-similarity
library_name: sentence-transformers
metrics:
  - pearson_cosine
  - spearman_cosine
model-index:
  - name: SentenceTransformer based on NbAiLab/nb-bert-large
    results:
      - task:
          type: semantic-similarity
          name: Semantic Similarity
        dataset:
          name: sts dev
          type: sts-dev
        metrics:
          - type: pearson_cosine
            value: 0.8522873287740148
            name: Pearson Cosine
          - type: spearman_cosine
            value: 0.8542943150375141
            name: Spearman Cosine

SentenceTransformer based on NbAiLab/nb-bert-large

This is a sentence-transformers model finetuned from NbAiLab/nb-bert-large. It builds on the previous work of the existing NbAiLab/nb-sbert-base model, using a larger foundational model and providing a larger max sequence length for inputs.

The model maps sentences & paragraphs to a 1024-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-large
Maximum Sequence Length: 512 tokens
Output Dimensionality: 1024 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': 1024, '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-large")
# Run inference
sentences = [
    "This is a Norwegian boy", 
    "Dette er en norsk gutt"
    ]

embeddings = model.encode(sentences)
print(embeddings.shape)
# (2, 1024)

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 0.9288],
#         [0.9288, 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-large')
model = AutoModel.from_pretrained('NbAiLab/nb-sbert-v2-large')

# 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, 1024])

similarity = cosine_similarity(embeddings[0].reshape(1, -1), embeddings[1].reshape(1, -1))
print(similarity)
# This should give 0.9288 in the example above.

Evaluation

Metrics

Semantic Similarity

Dataset: STS Benchmark
Evaluated with EmbeddingSimilarityEvaluator

Metric	nb-sbert-base	nb-sbert-v2-large
pearson_cosine	0.8275	0.8523
spearman_cosine	0.8245	0.8543

MTEB (Scandinavian)

Metric	nb-sbert-base	nb-sbert-v2-large
Mean (Task)	0.5190	0.5638
Mean (TaskType)	0.5394	0.5770

Bitext Mining	0.7228	0.7196
Classification	0.5708	0.6048
Clustering	0.3798	0.4005
Retrieval	0.4840	0.5832

Training Details

Training Dataset

Unnamed Dataset

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: 19.56 tokens max: 128 tokens	min: 5 tokens mean: 19.53 tokens max: 128 tokens	min: 6 tokens mean: 13.39 tokens max: 38 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.`	`Rommene er fine, men det offentlige området er i en egen liga.`	`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,
    "directions": [
        "query_to_doc"
    ],
    "partition_mode": "joint",
    "hardness_mode": null,
    "hardness_strength": 0.0
}

Training Hyperparameters

Non-Default Hyperparameters

per_device_train_batch_size: 32
gradient_accumulation_steps: 4
num_train_epochs: 1
learning_rate: 2e-05
warmup_steps: 412
bf16: True
batch_sampler: no_duplicates

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: steps
prediction_loss_only: True
per_device_train_batch_size: 32
per_device_eval_batch_size: 32
per_gpu_train_batch_size: None
per_gpu_eval_batch_size: None
gradient_accumulation_steps: 4
eval_accumulation_steps: None
torch_empty_cache_steps: None
learning_rate: 2e-05
weight_decay: 0.0
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1e-08
max_grad_norm: 1.0
num_train_epochs: 1
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: {}
warmup_ratio: 0.0
warmup_steps: 412
log_level: passive
log_level_replica: warning
log_on_each_node: True
logging_nan_inf_filter: True
save_safetensors: True
save_on_each_node: False
save_only_model: False
restore_callback_states_from_checkpoint: False
no_cuda: False
use_cpu: False
use_mps_device: False
seed: 42
data_seed: None
jit_mode_eval: False
bf16: True
fp16: False
fp16_opt_level: O1
half_precision_backend: auto
bf16_full_eval: False
fp16_full_eval: False
tf32: None
local_rank: 0
ddp_backend: None
tpu_num_cores: None
tpu_metrics_debug: False
debug: []
dataloader_drop_last: False
dataloader_num_workers: 0
dataloader_prefetch_factor: None
past_index: -1
disable_tqdm: False
remove_unused_columns: True
label_names: None
load_best_model_at_end: False
ignore_data_skip: False
fsdp: []
fsdp_min_num_params: 0
fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
fsdp_transformer_layer_cls_to_wrap: None
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
deepspeed: None
label_smoothing_factor: 0.0
optim: adamw_torch_fused
optim_args: None
adafactor: False
group_by_length: False
length_column_name: length
project: huggingface
trackio_space_id: trackio
ddp_find_unused_parameters: None
ddp_bucket_cap_mb: None
ddp_broadcast_buffers: False
dataloader_pin_memory: True
dataloader_persistent_workers: False
skip_memory_metrics: True
use_legacy_prediction_loop: False
push_to_hub: False
resume_from_checkpoint: None
hub_model_id: None
hub_strategy: every_save
hub_private_repo: None
hub_always_push: False
hub_revision: None
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
include_inputs_for_metrics: False
include_for_metrics: []
eval_do_concat_batches: True
fp16_backend: auto
push_to_hub_model_id: None
push_to_hub_organization: None
mp_parameters:
auto_find_batch_size: False
full_determinism: False
torchdynamo: None
ray_scope: last
ddp_timeout: 1800
torch_compile: False
torch_compile_backend: None
torch_compile_mode: None
include_tokens_per_second: False
include_num_input_tokens_seen: no
neftune_noise_alpha: None
optim_target_modules: None
batch_eval_metrics: False
eval_on_start: False
use_liger_kernel: False
liger_kernel_config: None
eval_use_gather_object: False
average_tokens_across_devices: True
prompts: None
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	sts-dev_spearman_cosine
0.0243	100	1.1675	-
0.0486	200	0.4302	-
0.0729	300	0.335	-
0.0971	400	0.2848	-
0.1000	412	-	0.8675
0.1214	500	0.2732	-
0.1457	600	0.26	-
0.1700	700	0.2433	-
0.1943	800	0.2281	-
0.2001	824	-	0.8612
0.2186	900	0.2381	-
0.2428	1000	0.218	-
0.2671	1100	0.2207	-
0.2914	1200	0.2163	-
0.3001	1236	-	0.8604
0.3157	1300	0.196	-
0.3400	1400	0.1924	-
0.3643	1500	0.1882	-
0.3885	1600	0.1923	-
0.4002	1648	-	0.8528
0.4128	1700	0.1796	-
0.4371	1800	0.1789	-
0.4614	1900	0.1935	-
0.4857	2000	0.1736	-
0.5002	2060	-	0.8614
0.5100	2100	0.1796	-
0.5342	2200	0.1826	-
0.5585	2300	0.1661	-
0.5828	2400	0.1676	-
0.6003	2472	-	0.8528
0.6071	2500	0.1631
0.6314	2600	0.1617	-
0.6557	2700	0.1652	-
0.6799	2800	0.1575	-
0.7003	2884	-	0.8490
0.7042	2900	0.1588	-
0.7285	3000	0.1535	-
0.7528	3100	0.1483	-
0.7771	3200	0.1463	-
0.8004	3296	-	0.8524
0.8014	3300	0.155
0.8256	3400	0.1472	-
0.8499	3500	0.153	-
0.8742	3600	0.1432	-
0.8985	3700	0.1467	-
0.9004	3708	-	0.8543
0.9228	3800	0.1473	-
0.9471	3900	0.1443	-
0.9713	4000	0.142	-
0.9956	4100	0.1396	-

Framework Versions

Python: 3.12.3
Sentence Transformers: 5.3.0
Transformers: 4.57.3
PyTorch: 2.9.1+rocm6.4
Accelerate: 1.12.0
Datasets: 4.6.0
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{oord2019representationlearningcontrastivepredictive,
      title={Representation Learning with Contrastive Predictive Coding},
      author={Aaron van den Oord and Yazhe Li and Oriol Vinyals},
      year={2019},
      eprint={1807.03748},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/1807.03748},
}

NbAiLab/nb-bert-large

@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.