metadata

base_model: BAAI/bge-m3
datasets: []
language:
  - es
library_name: sentence-transformers
license: apache-2.0
metrics:
  - cosine_accuracy@1
  - cosine_accuracy@3
  - cosine_accuracy@5
  - cosine_accuracy@10
  - cosine_precision@1
  - cosine_precision@3
  - cosine_precision@5
  - cosine_precision@10
  - cosine_recall@1
  - cosine_recall@3
  - cosine_recall@5
  - cosine_recall@10
  - cosine_ndcg@10
  - cosine_mrr@10
  - cosine_map@100
pipeline_tag: sentence-similarity
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:81
  - loss:MatryoshkaLoss
  - loss:MultipleNegativesRankingLoss
widget:
  - source_sentence: >-
      Disposeu del servei OAC360º d'assistència en la tramitació electrònica amb
      el que podeu contactar de dilluns a divendres de 08:00 a 20:00 a través
      del tel. 935 955 094, del correu oac360@sitges.cat, o del servei Truca'm,
      omplint el formulari perquè us truquin.
    sentences:
      - >-
        Com es pot demanar la comunicació prèvia d'obres per instal·lacions de
        plaques solars en sol urbà?
      - Quin és el correu electrònic per contactar amb el servei OAC360º?
      - Quin és l'efecte del silenci administratiu?
  - source_sentence: >-
      Positiu, llevat els casos en els quals manquin informes preceptius i
      vinculants d’altres administracions o d’aquells en els què es
      transfereixin al sol·licitant facultats contràries al planejament i la
      legislació urbanística.
    sentences:
      - Quin és el document que cal aportar per a aquest tràmit?
      - >-
        Quin és el lloc on es pot tramitar la presentació de justificants de
        pagament per als ajuts del lloguer just dels habitatges?
      - >-
        Quin és el sentit del silenci administratiu per a la comunicació prèvia
        d'obres per instal·lacions de plaques solars en sol urbà?
model-index:
  - name: BGE large Legal Spanish
    results:
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 1024
          type: dim_1024
        metrics:
          - type: cosine_accuracy@1
            value: 0.1111111111111111
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.3333333333333333
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.4444444444444444
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.7777777777777778
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.1111111111111111
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.1111111111111111
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.08888888888888889
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.07777777777777778
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.1111111111111111
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.3333333333333333
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.4444444444444444
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.7777777777777778
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.37561164042849293
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.2550705467372134
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.26453109424123916
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 768
          type: dim_768
        metrics:
          - type: cosine_accuracy@1
            value: 0.1111111111111111
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.3333333333333333
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.4444444444444444
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.7777777777777778
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.1111111111111111
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.1111111111111111
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.08888888888888889
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.07777777777777778
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.1111111111111111
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.3333333333333333
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.4444444444444444
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.7777777777777778
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.37561164042849293
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.2550705467372134
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.26591710758377424
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 512
          type: dim_512
        metrics:
          - type: cosine_accuracy@1
            value: 0.1111111111111111
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.3333333333333333
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.4444444444444444
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.7777777777777778
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.1111111111111111
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.1111111111111111
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.08888888888888889
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.07777777777777778
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.1111111111111111
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.3333333333333333
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.4444444444444444
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.7777777777777778
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.36941287151905455
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.24828042328042324
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.25912698412698415
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 256
          type: dim_256
        metrics:
          - type: cosine_accuracy@1
            value: 0.1111111111111111
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.3333333333333333
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.4444444444444444
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.6666666666666666
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.1111111111111111
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.1111111111111111
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.08888888888888889
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.06666666666666668
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.1111111111111111
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.3333333333333333
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.4444444444444444
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.6666666666666666
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.33724514013077883
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.23796296296296296
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.2553057025279247
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 128
          type: dim_128
        metrics:
          - type: cosine_accuracy@1
            value: 0.1111111111111111
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.3333333333333333
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.5555555555555556
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.7777777777777778
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.1111111111111111
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.1111111111111111
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.1111111111111111
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.07777777777777778
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.1111111111111111
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.3333333333333333
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.5555555555555556
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.7777777777777778
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.3920021980903836
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.27248677248677244
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.2795432240996757
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 64
          type: dim_64
        metrics:
          - type: cosine_accuracy@1
            value: 0.2222222222222222
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.3333333333333333
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.4444444444444444
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.5555555555555556
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.2222222222222222
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.1111111111111111
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.08888888888888889
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.05555555555555555
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.2222222222222222
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.3333333333333333
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.4444444444444444
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.5555555555555556
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.3626677657118585
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.3029100529100529
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.32598958775429365
            name: Cosine Map@100

BGE large Legal Spanish

This is a sentence-transformers model finetuned from BAAI/bge-m3. It 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.

Model Details

Model Description

Model Type: Sentence Transformer
Base model: BAAI/bge-m3
Maximum Sequence Length: 8192 tokens
Output Dimensionality: 1024 tokens
Similarity Function: Cosine Similarity
Language: es
License: apache-2.0

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': 8192, 'do_lower_case': False}) with Transformer model: XLMRobertaModel 
  (1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
  (2): Normalize()
)

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("adriansanz/bge-m3-es-legal-tmp-6")
# Run inference
sentences = [
    'Positiu, llevat els casos en els quals manquin informes preceptius i vinculants d’altres administracions o d’aquells en els què es transfereixin al sol·licitant facultats contràries al planejament i la legislació urbanística.',
    "Quin és el sentit del silenci administratiu per a la comunicació prèvia d'obres per instal·lacions de plaques solars en sol urbà?",
    'Quin és el lloc on es pot tramitar la presentació de justificants de pagament per als ajuts del lloguer just dels habitatges?',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]

Evaluation

Metrics

Information Retrieval

Dataset: dim_1024
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.1111
cosine_accuracy@3	0.3333
cosine_accuracy@5	0.4444
cosine_accuracy@10	0.7778
cosine_precision@1	0.1111
cosine_precision@3	0.1111
cosine_precision@5	0.0889
cosine_precision@10	0.0778
cosine_recall@1	0.1111
cosine_recall@3	0.3333
cosine_recall@5	0.4444
cosine_recall@10	0.7778
cosine_ndcg@10	0.3756
cosine_mrr@10	0.2551
cosine_map@100	0.2645

Information Retrieval

Dataset: dim_768
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.1111
cosine_accuracy@3	0.3333
cosine_accuracy@5	0.4444
cosine_accuracy@10	0.7778
cosine_precision@1	0.1111
cosine_precision@3	0.1111
cosine_precision@5	0.0889
cosine_precision@10	0.0778
cosine_recall@1	0.1111
cosine_recall@3	0.3333
cosine_recall@5	0.4444
cosine_recall@10	0.7778
cosine_ndcg@10	0.3756
cosine_mrr@10	0.2551
cosine_map@100	0.2659

Information Retrieval

Dataset: dim_512
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.1111
cosine_accuracy@3	0.3333
cosine_accuracy@5	0.4444
cosine_accuracy@10	0.7778
cosine_precision@1	0.1111
cosine_precision@3	0.1111
cosine_precision@5	0.0889
cosine_precision@10	0.0778
cosine_recall@1	0.1111
cosine_recall@3	0.3333
cosine_recall@5	0.4444
cosine_recall@10	0.7778
cosine_ndcg@10	0.3694
cosine_mrr@10	0.2483
cosine_map@100	0.2591

Information Retrieval

Dataset: dim_256
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.1111
cosine_accuracy@3	0.3333
cosine_accuracy@5	0.4444
cosine_accuracy@10	0.6667
cosine_precision@1	0.1111
cosine_precision@3	0.1111
cosine_precision@5	0.0889
cosine_precision@10	0.0667
cosine_recall@1	0.1111
cosine_recall@3	0.3333
cosine_recall@5	0.4444
cosine_recall@10	0.6667
cosine_ndcg@10	0.3372
cosine_mrr@10	0.238
cosine_map@100	0.2553

Information Retrieval

Dataset: dim_128
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.1111
cosine_accuracy@3	0.3333
cosine_accuracy@5	0.5556
cosine_accuracy@10	0.7778
cosine_precision@1	0.1111
cosine_precision@3	0.1111
cosine_precision@5	0.1111
cosine_precision@10	0.0778
cosine_recall@1	0.1111
cosine_recall@3	0.3333
cosine_recall@5	0.5556
cosine_recall@10	0.7778
cosine_ndcg@10	0.392
cosine_mrr@10	0.2725
cosine_map@100	0.2795

Information Retrieval

Dataset: dim_64
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.2222
cosine_accuracy@3	0.3333
cosine_accuracy@5	0.4444
cosine_accuracy@10	0.5556
cosine_precision@1	0.2222
cosine_precision@3	0.1111
cosine_precision@5	0.0889
cosine_precision@10	0.0556
cosine_recall@1	0.2222
cosine_recall@3	0.3333
cosine_recall@5	0.4444
cosine_recall@10	0.5556
cosine_ndcg@10	0.3627
cosine_mrr@10	0.3029
cosine_map@100	0.326

Training Details

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: epoch
per_device_train_batch_size: 16
per_device_eval_batch_size: 16
gradient_accumulation_steps: 16
learning_rate: 2e-05
num_train_epochs: 6
lr_scheduler_type: cosine
warmup_ratio: 0.1
bf16: True
tf32: False
load_best_model_at_end: True
optim: adamw_torch_fused
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: 16
per_device_eval_batch_size: 16
per_gpu_train_batch_size: None
per_gpu_eval_batch_size: None
gradient_accumulation_steps: 16
eval_accumulation_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: 6
max_steps: -1
lr_scheduler_type: cosine
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: True
fp16: False
fp16_opt_level: O1
half_precision_backend: auto
bf16_full_eval: False
fp16_full_eval: False
tf32: False
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: True
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}
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: False
hub_always_push: False
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
include_inputs_for_metrics: False
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
dispatch_batches: None
split_batches: 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
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	loss	dim_1024_cosine_map@100	dim_128_cosine_map@100	dim_256_cosine_map@100	dim_512_cosine_map@100	dim_64_cosine_map@100	dim_768_cosine_map@100
1.0	1	-	3.7675	0.2475	0.2919	0.2372	0.2511	0.2510	0.2468
2.0	2	-	3.9701	0.2533	0.3028	0.2473	0.2601	0.3449	0.2716
3.0	4	-	4.1211	0.2645	0.2704	0.2548	0.2614	0.3283	0.2654
4.0	5	1.8359	4.0228	0.2645	0.2789	0.2553	0.2619	0.3260	0.2659
5.0	6	-	3.9758	0.2645	0.2795	0.2553	0.2591	0.3260	0.2659

The bold row denotes the saved checkpoint.

Framework Versions

Python: 3.10.12
Sentence Transformers: 3.0.1
Transformers: 4.42.3
PyTorch: 2.3.1+cu121
Accelerate: 0.32.1
Datasets: 2.20.0
Tokenizers: 0.19.1

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",
}

MatryoshkaLoss

@misc{kusupati2024matryoshka,
    title={Matryoshka Representation Learning}, 
    author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
    year={2024},
    eprint={2205.13147},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

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}
}