This is a sentence-transformers model finetuned from BAAI/bge-m3 on the json dataset. 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
Training Dataset:
- json

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.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("adriansanz/ST-tramits-SB-001-5ep")
# Run inference
sentences = [
    'Descripció. Retorna en format JSON adequat',
    "Quin és el contingut de l'annex específic?",
    "Què passa amb l'habitatge?",
]
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	0.3322
cosine_accuracy@3	0.5902
cosine_accuracy@5	0.6998
cosine_accuracy@10	0.8094
cosine_precision@1	0.3322
cosine_precision@3	0.1967
cosine_precision@5	0.14
cosine_precision@10	0.0809
cosine_recall@1	0.3322
cosine_recall@3	0.5902
cosine_recall@5	0.6998
cosine_recall@10	0.8094
cosine_ndcg@10	0.5626
cosine_mrr@10	0.4843
cosine_map@100	0.4924

Information Retrieval

Dataset: dim_768
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.3406
cosine_accuracy@3	0.5767
cosine_accuracy@5	0.6981
cosine_accuracy@10	0.8162
cosine_precision@1	0.3406
cosine_precision@3	0.1922
cosine_precision@5	0.1396
cosine_precision@10	0.0816
cosine_recall@1	0.3406
cosine_recall@3	0.5767
cosine_recall@5	0.6981
cosine_recall@10	0.8162
cosine_ndcg@10	0.5661
cosine_mrr@10	0.4872
cosine_map@100	0.4952

Information Retrieval

Dataset: dim_512
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.3305
cosine_accuracy@3	0.5801
cosine_accuracy@5	0.6948
cosine_accuracy@10	0.8162
cosine_precision@1	0.3305
cosine_precision@3	0.1934
cosine_precision@5	0.139
cosine_precision@10	0.0816
cosine_recall@1	0.3305
cosine_recall@3	0.5801
cosine_recall@5	0.6948
cosine_recall@10	0.8162
cosine_ndcg@10	0.563
cosine_mrr@10	0.483
cosine_map@100	0.4908

Information Retrieval

Dataset: dim_256
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.3288
cosine_accuracy@3	0.5885
cosine_accuracy@5	0.7015
cosine_accuracy@10	0.8094
cosine_precision@1	0.3288
cosine_precision@3	0.1962
cosine_precision@5	0.1403
cosine_precision@10	0.0809
cosine_recall@1	0.3288
cosine_recall@3	0.5885
cosine_recall@5	0.7015
cosine_recall@10	0.8094
cosine_ndcg@10	0.5626
cosine_mrr@10	0.4842
cosine_map@100	0.492

Information Retrieval

Dataset: dim_128
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.3474
cosine_accuracy@3	0.5818
cosine_accuracy@5	0.6998
cosine_accuracy@10	0.8061
cosine_precision@1	0.3474
cosine_precision@3	0.1939
cosine_precision@5	0.14
cosine_precision@10	0.0806
cosine_recall@1	0.3474
cosine_recall@3	0.5818
cosine_recall@5	0.6998
cosine_recall@10	0.8061
cosine_ndcg@10	0.5654
cosine_mrr@10	0.4894
cosine_map@100	0.4973

Information Retrieval

Dataset: dim_64
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.2917
cosine_accuracy@3	0.5683
cosine_accuracy@5	0.6644
cosine_accuracy@10	0.7875
cosine_precision@1	0.2917
cosine_precision@3	0.1894
cosine_precision@5	0.1329
cosine_precision@10	0.0788
cosine_recall@1	0.2917
cosine_recall@3	0.5683
cosine_recall@5	0.6644
cosine_recall@10	0.7875
cosine_ndcg@10	0.532
cosine_mrr@10	0.4512
cosine_map@100	0.4595

Training Details

Training Dataset

json

Dataset: json
Size: 2,372 training samples
Columns: positive and anchor
Approximate statistics based on the first 1000 samples:
positive anchor
type string string
details
min: 3 tokens
mean: 35.12 tokens
max: 166 tokens
min: 8 tokens
mean: 19.49 tokens
max: 47 tokens

Samples:

positive	anchor
`Comunicar la variació d'alguna de les següents dades del Padró Municipal d'Habitants: Nom, Cognoms, Data de naixement, DNI, Passaport, Número de permís de residència (NIE), Sexe, Municipi i/o província de naixement, Nacionalitat, Titulació acadèmica.`	`Quin és l'objectiu del canvi de dades personals en el Padró Municipal d'Habitants?`
`EN QUÈ CONSISTEIX: Tramitar la sol·licitud de matrimoni civil a l'Ajuntament.`	`Què és el matrimoni civil a l'Ajuntament de Sant Boi de Llobregat?`
`En domiciliar el pagament de tributs municipals en entitats bancàries.`	`Quin és el benefici de domiciliar el pagament de tributs?`

Loss: MatryoshkaLoss with these parameters:
json
1{ 2 "loss": "MultipleNegativesRankingLoss", 3 "matryoshka_dims": [ 4 1024, 5 768, 6 512, 7 256, 8 128, 9 64 10 ], 11 "matryoshka_weights": [ 12 1, 13 1, 14 1, 15 1, 16 1, 17 1 18 ], 19 "n_dims_per_step": -1 20}

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: 5
lr_scheduler_type: cosine
warmup_ratio: 0.2
bf16: True
tf32: True
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
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: 5
max_steps: -1
lr_scheduler_type: cosine
lr_scheduler_kwargs: {}
warmup_ratio: 0.2
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: True
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
eval_use_gather_object: False
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	dim_1024_cosine_map@100	dim_768_cosine_map@100	dim_512_cosine_map@100	dim_256_cosine_map@100	dim_128_cosine_map@100	dim_64_cosine_map@100
0.9664	9	-	0.4730	0.4766	0.4640	0.4612	0.4456	0.4083
1.0738	10	2.6023	-	-	-	-	-	-
1.9329	18	-	0.4951	0.4966	0.4977	0.4773	0.4849	0.4501
2.1477	20	0.974	-	-	-	-	-	-
2.8993	27	-	0.4891	0.4973	0.4941	0.4867	0.4925	0.4684
3.2215	30	0.408	-	-	-	-	-	-
3.9732	37	-	0.4944	0.4998	0.4931	0.4991	0.4974	0.4616
4.2953	40	0.2718	-	-	-	-	-	-
4.8322	45	-	0.4924	0.4952	0.4908	0.4920	0.4973	0.4595

The bold row denotes the saved checkpoint.

Framework Versions

Python: 3.10.12
Sentence Transformers: 3.2.0
Transformers: 4.44.2
PyTorch: 2.4.1+cu121
Accelerate: 1.1.0.dev0
Datasets: 3.0.1
Tokenizers: 0.19.1

Citation

BibTeX

Sentence Transformers

bibtex

@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

bibtex

@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

bibtex

@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}
}

	positive	anchor
type	string	string
details	min: 3 tokens mean: 35.12 tokens max: 166 tokens	min: 8 tokens mean: 19.49 tokens max: 47 tokens