SentenceTransformer based on HiTZ/BERnaT_base
This is a sentence-transformers model finetuned from HiTZ/BERnaT_base. It 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.
Model Details
Model Description
- Model Type: Sentence Transformer
- Base model: HiTZ/BERnaT_base
- Maximum Sequence Length: 512 tokens
- Output Dimensionality: 768 dimensions
- Similarity Function: Cosine Similarity
Model Sources
Full Model Architecture
SentenceTransformer(
(0): Transformer({'max_seq_length': 512, 'do_lower_case': False, 'architecture': 'RobertaModel'})
(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
model = SentenceTransformer("aimarsg/bernat_parl_l_contrastive")
sentences = [
'Noiz egin zen itunpeko irakaskuntzaren finantzaketari buruzko barne-dokumentua?',
'[TOPIC: Galdera, José Manuel Gil Vegas Euskal Talde Popularra-Ciudadanos taldeko legebiltzarkideak Hezkuntzako sailburuari egina, itunpeko irakaskuntzaren finantzaketari buruz]\n[GIL VEGAS, (Mixto-Ciudadanos)]:\nEskerrik asko erantzunagatik, sailburu jauna. Hala ere, ez didazu gauza handirik argitu. Barne-dokumentu bat dela esan didazu, 2021eko apirilean egin zena, baina oraindik ez dakigu zergatik duen prentsak Ganbera honek baino lehenago. Eta esan didazu ikastetxeei eman zitzaiela, baina ez zela Legebiltzar honetara bidali. Eta, beraz, uler ezazu galdetzen jarraitu beharko dudala txosten hori gaur egun non dagoen eta zein den hark duen estatusa; galdetu ez ezik, berriro diot, (Date: 18.02.2022)',
'Gogoan izan behar da dokumentu hori ez dela inolako akordioren ondorio, ez negoziazioen ondorio, ezta horri buruzko elkarrizketen ondorio ere, presidentearen ikuspuntua besterik ez baita.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
similarities = model.similarity(embeddings, embeddings)
print(similarities)
Evaluation
Metrics
Triplet
| Metric |
Value |
| cosine_accuracy |
0.9059 |
Training Details
Training Dataset
Unnamed Dataset
Evaluation Dataset
Unnamed Dataset
Training Hyperparameters
Non-Default Hyperparameters
eval_strategy: epoch
learning_rate: 2e-05
warmup_ratio: 0.1
batch_sampler: no_duplicates
All Hyperparameters
Click to expand
overwrite_output_dir: False
do_predict: False
eval_strategy: epoch
prediction_loss_only: True
per_device_train_batch_size: 8
per_device_eval_batch_size: 8
per_gpu_train_batch_size: None
per_gpu_eval_batch_size: None
gradient_accumulation_steps: 1
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: 3
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: {}
warmup_ratio: 0.1
warmup_steps: 0
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
use_ipex: False
bf16: False
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
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: False
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: False
prompts: None
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional
router_mapping: {}
learning_rate_mapping: {}
Training Logs
| Epoch |
Step |
Training Loss |
Validation Loss |
multilingual-e5-large_cosine_accuracy |
| 0.0666 |
100 |
1.5709 |
- |
- |
| 0.1332 |
200 |
0.2662 |
- |
- |
| 0.1999 |
300 |
0.148 |
- |
- |
| 0.2665 |
400 |
0.1345 |
- |
- |
| 0.3331 |
500 |
0.1064 |
- |
- |
| 0.3997 |
600 |
0.1244 |
- |
- |
| 0.4664 |
700 |
0.0665 |
- |
- |
| 0.5330 |
800 |
0.0825 |
- |
- |
| 0.5996 |
900 |
0.0768 |
- |
- |
| 0.6662 |
1000 |
0.0741 |
- |
- |
| 0.7328 |
1100 |
0.0496 |
- |
- |
| 0.7995 |
1200 |
0.069 |
- |
- |
| 0.8661 |
1300 |
0.0814 |
- |
- |
| 0.9327 |
1400 |
0.0744 |
- |
- |
| 0.9993 |
1500 |
0.0536 |
- |
- |
| 1.0 |
1501 |
- |
0.3427 |
0.8788 |
| 1.0660 |
1600 |
0.029 |
- |
- |
| 1.1326 |
1700 |
0.0375 |
- |
- |
| 1.1992 |
1800 |
0.0332 |
- |
- |
| 1.2658 |
1900 |
0.027 |
- |
- |
| 1.3324 |
2000 |
0.033 |
- |
- |
| 1.3991 |
2100 |
0.0441 |
- |
- |
| 1.4657 |
2200 |
0.0261 |
- |
- |
| 1.5323 |
2300 |
0.0278 |
- |
- |
| 1.5989 |
2400 |
0.0217 |
- |
- |
| 1.6656 |
2500 |
0.0126 |
- |
- |
| 1.7322 |
2600 |
0.0139 |
- |
- |
| 1.7988 |
2700 |
0.0325 |
- |
- |
| 1.8654 |
2800 |
0.0233 |
- |
- |
| 1.9320 |
2900 |
0.0311 |
- |
- |
| 1.9987 |
3000 |
0.0144 |
- |
- |
| 2.0 |
3002 |
- |
0.2501 |
0.9047 |
| 2.0653 |
3100 |
0.0069 |
- |
- |
| 2.1319 |
3200 |
0.0123 |
- |
- |
| 2.1985 |
3300 |
0.0094 |
- |
- |
| 2.2652 |
3400 |
0.0184 |
- |
- |
| 2.3318 |
3500 |
0.0109 |
- |
- |
| 2.3984 |
3600 |
0.0083 |
- |
- |
| 2.4650 |
3700 |
0.0158 |
- |
- |
| 2.5316 |
3800 |
0.0169 |
- |
- |
| 2.5983 |
3900 |
0.0137 |
- |
- |
| 2.6649 |
4000 |
0.0069 |
- |
- |
| 2.7315 |
4100 |
0.0063 |
- |
- |
| 2.7981 |
4200 |
0.0102 |
- |
- |
| 2.8648 |
4300 |
0.0107 |
- |
- |
| 2.9314 |
4400 |
0.0083 |
- |
- |
| 2.9980 |
4500 |
0.0188 |
- |
- |
| 3.0 |
4503 |
- |
0.2361 |
0.9059 |
Framework Versions
- Python: 3.10.8
- Sentence Transformers: 5.1.0
- Transformers: 4.56.0
- PyTorch: 2.8.0+cu128
- Accelerate: 1.10.1
- Datasets: 4.0.0
- Tokenizers: 0.22.0
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}
}