embed-andegpt-H384

This is a sentence-transformers model finetuned from nreimers/MiniLM-L6-H384-uncased. It maps sentences & paragraphs to a 384-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: nreimers/MiniLM-L6-H384-uncased
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 384 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': 512, 'do_lower_case': False}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 384, '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("enpaiva/embed-andegpt-H384")
# Run inference
sentences = [
    '¿Cuál es el nombre del reglamento que se menciona en la información proporcionada?',
    'Reglamento de Baja Tensión de la ANDE: El 10- trata sobre Partes de que se compone una instalación eléctrica: y tiene las siguientes sub-secciones: <sub-section>10.1</sub-section>',
    'Reglamento de Baja Tensión de la ANDE: El 37- trata sobre Soldadura eléctrica: y tiene las siguientes sub-secciones: <sub-section>37.1</sub-section>, <sub-section>37.2</sub-section>, <sub-section>37.3</sub-section>, <sub-section>37.4</sub-section>',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]

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

Evaluation

Metrics

Triplet

• Dataset: andegpt-dev
• Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.9983
dot_accuracy	0.0022
manhattan_accuracy	0.9985
euclidean_accuracy	0.9983
max_accuracy	0.9985

Triplet

• Dataset: andegpt-test
• Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.9973
dot_accuracy	0.0024
manhattan_accuracy	0.9971
euclidean_accuracy	0.9973
max_accuracy	0.9973

Training Details

Training Hyperparameters

Non-Default Hyperparameters

• prediction_loss_only: False
• per_device_train_batch_size: 32
• learning_rate: 2e-05
• lr_scheduler_type: cosine
• log_level_replica: passive
• log_on_each_node: False
• logging_nan_inf_filter: False
• bf16: True
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• prediction_loss_only: False
• per_device_train_batch_size: 32
• 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
• 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: cosine
• lr_scheduler_kwargs: {}
• warmup_ratio: 0
• warmup_steps: 0
• log_level: passive
• log_level_replica: passive
• log_on_each_node: False
• logging_nan_inf_filter: False
• save_safetensors: True
• save_on_each_node: False
• save_only_model: 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: 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}
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch
• 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
• 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_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	loss	andegpt-dev_max_accuracy	andegpt-test_max_accuracy
0	0	-	-	0.5920	-
0.1079	250	2.3094	0.7200	0.9597	-
0.2158	500	0.7952	0.3598	0.9813	-
0.3237	750	0.4862	0.2162	0.9910	-
0.4316	1000	0.3304	0.1558	0.9927	-
0.5395	1250	0.2527	0.1140	0.9961	-
0.6474	1500	0.1987	0.0859	0.9964	-
0.7553	1750	0.1617	0.0729	0.9959	-
0.8632	2000	0.1419	0.0562	0.9966	-
0.9711	2250	0.1132	0.0495	0.9968	-
1.0790	2500	0.1043	0.0429	0.9971	-
1.1869	2750	0.0947	0.0368	0.9978	-
1.2948	3000	0.0736	0.0367	0.9976	-
1.4027	3250	0.0661	0.0296	0.9978	-
1.5106	3500	0.0613	0.0279	0.9985	-
1.6185	3750	0.0607	0.0264	0.9983	-
1.7264	4000	0.0521	0.0238	0.9985	-
1.8343	4250	0.0495	0.0216	0.9985	-
1.9422	4500	0.0425	0.0211	0.9983	-
2.0501	4750	0.0428	0.0200	0.9983	-
2.1580	5000	0.0435	0.0190	0.9985	-
2.2659	5250	0.0393	0.0188	0.9983	-
2.3738	5500	0.0356	0.0182	0.9983	-
2.4817	5750	0.0351	0.0180	0.9988	-
2.5896	6000	0.0394	0.0181	0.9985	-
2.5973	6018	-	-	-	0.9973

Framework Versions

• Python: 3.11.0
• Sentence Transformers: 3.0.1
• Transformers: 4.39.3
• PyTorch: 2.2.0+cu121
• Accelerate: 0.28.0
• Datasets: 2.20.0
• Tokenizers: 0.15.2

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{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}
}