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-MONT-001-5ep")
# Run inference
sentences = [
    "Laboratori d'art: D'octubre 2024 a maig de 2025. Horari: Dilluns de 17.30h a 19.00h",
    "Quin és el dia i hora del curs de Laboratori d'art?",
    "On es pot sol·licitar la reserva especial d'estacionament?",
]
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	0.1038
cosine_accuracy@3	0.2154
cosine_accuracy@5	0.2769
cosine_accuracy@10	0.4885
cosine_precision@1	0.1038
cosine_precision@3	0.0718
cosine_precision@5	0.0554
cosine_precision@10	0.0488
cosine_recall@1	0.1038
cosine_recall@3	0.2154
cosine_recall@5	0.2769
cosine_recall@10	0.4885
cosine_ndcg@10	0.2612
cosine_mrr@10	0.1933
cosine_map@100	0.2192

Information Retrieval

Dataset: dim_768
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.1192
cosine_accuracy@3	0.2308
cosine_accuracy@5	0.3115
cosine_accuracy@10	0.5308
cosine_precision@1	0.1192
cosine_precision@3	0.0769
cosine_precision@5	0.0623
cosine_precision@10	0.0531
cosine_recall@1	0.1192
cosine_recall@3	0.2308
cosine_recall@5	0.3115
cosine_recall@10	0.5308
cosine_ndcg@10	0.2878
cosine_mrr@10	0.215
cosine_map@100	0.2378

Information Retrieval

Dataset: dim_512
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.1269
cosine_accuracy@3	0.2385
cosine_accuracy@5	0.3269
cosine_accuracy@10	0.5269
cosine_precision@1	0.1269
cosine_precision@3	0.0795
cosine_precision@5	0.0654
cosine_precision@10	0.0527
cosine_recall@1	0.1269
cosine_recall@3	0.2385
cosine_recall@5	0.3269
cosine_recall@10	0.5269
cosine_ndcg@10	0.292
cosine_mrr@10	0.2216
cosine_map@100	0.2444

Information Retrieval

Dataset: dim_256
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.1038
cosine_accuracy@3	0.2077
cosine_accuracy@5	0.3077
cosine_accuracy@10	0.4962
cosine_precision@1	0.1038
cosine_precision@3	0.0692
cosine_precision@5	0.0615
cosine_precision@10	0.0496
cosine_recall@1	0.1038
cosine_recall@3	0.2077
cosine_recall@5	0.3077
cosine_recall@10	0.4962
cosine_ndcg@10	0.2649
cosine_mrr@10	0.1953
cosine_map@100	0.2202

Information Retrieval

Dataset: dim_128
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.1308
cosine_accuracy@3	0.2308
cosine_accuracy@5	0.3423
cosine_accuracy@10	0.55
cosine_precision@1	0.1308
cosine_precision@3	0.0769
cosine_precision@5	0.0685
cosine_precision@10	0.055
cosine_recall@1	0.1308
cosine_recall@3	0.2308
cosine_recall@5	0.3423
cosine_recall@10	0.55
cosine_ndcg@10	0.3001
cosine_mrr@10	0.2253
cosine_map@100	0.2489

Information Retrieval

Dataset: dim_64
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.1038
cosine_accuracy@3	0.2231
cosine_accuracy@5	0.2846
cosine_accuracy@10	0.4923
cosine_precision@1	0.1038
cosine_precision@3	0.0744
cosine_precision@5	0.0569
cosine_precision@10	0.0492
cosine_recall@1	0.1038
cosine_recall@3	0.2231
cosine_recall@5	0.2846
cosine_recall@10	0.4923
cosine_ndcg@10	0.2636
cosine_mrr@10	0.195
cosine_map@100	0.2197

Training Details

Training Dataset

json

Dataset: json
Size: 3,814 training samples
Columns: positive and anchor
Approximate statistics based on the first 1000 samples:
positive anchor
type string string
details
min: 3 tokens
mean: 39.27 tokens
max: 165 tokens
min: 9 tokens
mean: 20.67 tokens
max: 50 tokens

Samples:

positive	anchor
`Aquest tràmit permet obtenir la llicència per a ocupació de la via pública per quioscs, casetes o parades (xurreries, gelats,...).`	`Quins són els requisits per obtenir la llicència d'ocupació de la via pública per a gelats?`
`Aquest tràmit permet obtenir la llicència d'ocupació de la via pública per a la instal·lació de grues desmuntables.`	`Quin és el lloc on es pot obtenir la llicència d'ocupació de la via pública per a la instal·lació de grues desmuntables en una via pública?`
`L’Espai Jove de Montgat disposa de dues sales, una aula, i una sala chill-out així com jardins i serveis adreçats als joves del municipi.`	`Quin és el propòsit de l'aula de l'Espai Jove de Montgat?`

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
0.6695	10	3.4242	-	-	-	-	-	-
0.9372	14	-	0.2075	0.2165	0.2078	0.1957	0.2050	0.1949
1.3389	20	1.666	-	-	-	-	-	-
1.9414	29	-	0.2145	0.2184	0.2248	0.2144	0.2244	0.2112
2.0084	30	0.7666	-	-	-	-	-	-
2.6778	40	0.4859	-	-	-	-	-	-
2.9456	44	-	0.2263	0.2408	0.2234	0.2274	0.252	0.2313
3.3473	50	0.277	-	-	-	-	-	-
3.9498	59	-	0.2107	0.2359	0.2386	0.2275	0.2382	0.2246
4.0167	60	0.2423	-	-	-	-	-	-
4.6862	70	0.2281	0.2192	0.2378	0.2444	0.2202	0.2489	0.2197

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: 39.27 tokens max: 165 tokens	min: 9 tokens mean: 20.67 tokens max: 50 tokens

adriansanz/ST-tramits-MONTGAT-001-5ep

SentenceTransformer based on BAAI/bge-m3

Model Details

Model Description

Model Sources

Full Model Architecture

Usage

Direct Usage (Sentence Transformers)

Evaluation

Metrics

Information Retrieval

Information Retrieval

Information Retrieval

Information Retrieval

Information Retrieval

Information Retrieval

Training Details

Training Dataset

json

Training Hyperparameters

Non-Default Hyperparameters

All Hyperparameters

Training Logs

Framework Versions

Citation

BibTeX

Sentence Transformers

MatryoshkaLoss

MultipleNegativesRankingLoss

adriansanz/
ST-tramits-MONTGAT-001-5ep