winderfeld/cc-uffs-ppc-ft-test-multiqa

SentenceTransformer based on sentence-transformers/multi-qa-mpnet-base-dot-v1

This is a sentence-transformers model finetuned from sentence-transformers/multi-qa-mpnet-base-dot-v1. 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: sentence-transformers/multi-qa-mpnet-base-dot-v1
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 tokens
• Similarity Function: Dot Product

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: MPNetModel 
  (1): Pooling({'word_embedding_dimension': 768, '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})
)

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("sentence_transformers_model_id")
# Run inference
sentences = [
    'Onde é mencionado oficialmente o NDE do curso de Ciência Da Computação, conforme a Portaria nº ?',
    '**3.4 Núcleo Docente Estruturante do Curso**<br><br>O NDE do curso de Ciência Da Computação, conforme designado na Portaria nº <br><br>Projeto Pedagógico do Curso de Ciência Da Computação,*Campus*Chapecó. <br><br>17 ',
    '**IDENTIFICAÇÃO INSTITUCIONAL**<br><br>A Universidade Federal da Fronteira Sul foi criada pela Lei Nº 12.029, de 15 de <br><br>35  setembro de 2009. Tem abrangência interestadual com sede na cidade catarinense de <br><br>Chapecó, três*campi*no Rio Grande do Sul – Cerro Largo, Erechim e Passo Fundo – e dois <br><br>*campi*no Paraná – Laranjeiras do Sul e Realeza. ',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

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

Evaluation

Metrics

Information Retrieval

• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.625
cosine_accuracy@3	0.8015
cosine_accuracy@5	0.864
cosine_accuracy@10	0.9228
cosine_precision@1	0.625
cosine_precision@3	0.2672
cosine_precision@5	0.1728
cosine_precision@10	0.0923
cosine_recall@1	0.625
cosine_recall@3	0.8015
cosine_recall@5	0.864
cosine_recall@10	0.9228
cosine_ndcg@10	0.7746
cosine_mrr@10	0.7271
cosine_map@100	0.7301
dot_accuracy@1	0.6275
dot_accuracy@3	0.799
dot_accuracy@5	0.8701
dot_accuracy@10	0.9203
dot_precision@1	0.6275
dot_precision@3	0.2663
dot_precision@5	0.174
dot_precision@10	0.092
dot_recall@1	0.6275
dot_recall@3	0.799
dot_recall@5	0.8701
dot_recall@10	0.9203
dot_ndcg@10	0.774
dot_mrr@10	0.7269
dot_map@100	0.7302

Training Details

Training Dataset

Unnamed Dataset

• Size: 2,012 training samples
• Columns: sentence_0 and sentence_1
• Approximate statistics based on the first 1000 samples:

  	sentence_0	sentence_1
  type	string	string
  details	min: 14 tokens mean: 40.7 tokens max: 123 tokens	min: 9 tokens mean: 272.17 tokens max: 512 tokens

• Samples:

  sentence_0	sentence_1
  Em quantos estados brasileiros a Universidade Federal da Fronteira Sul está localizada?	IDENTIFICAÇÃO INSTITUCIONAL A Universidade Federal da Fronteira Sul foi criada pela Lei Nº 12.029, de 15 de 35 setembro de 2009. Tem abrangência interestadual com sede na cidade catarinense de Chapecó, trêscampino Rio Grande do Sul – Cerro Largo, Erechim e Passo Fundo – e dois campino Paraná – Laranjeiras do Sul e Realeza.
  Qual é a cidade sede da universidade?	IDENTIFICAÇÃO INSTITUCIONAL A Universidade Federal da Fronteira Sul foi criada pela Lei Nº 12.029, de 15 de 35 setembro de 2009. Tem abrangência interestadual com sede na cidade catarinense de Chapecó, trêscampino Rio Grande do Sul – Cerro Largo, Erechim e Passo Fundo – e dois campino Paraná – Laranjeiras do Sul e Realeza.
  Quantos campi possui a universidade em cada um dos estados onde está presente?	IDENTIFICAÇÃO INSTITUCIONAL A Universidade Federal da Fronteira Sul foi criada pela Lei Nº 12.029, de 15 de 35 setembro de 2009. Tem abrangência interestadual com sede na cidade catarinense de Chapecó, trêscampino Rio Grande do Sul – Cerro Largo, Erechim e Passo Fundo – e dois campino Paraná – Laranjeiras do Sul e Realeza.

• Loss: MatryoshkaLoss with these parameters:

  {
      "loss": "MultipleNegativesRankingLoss",
      "matryoshka_dims": [
          768,
          512,
          256,
          128,
          64
      ],
      "matryoshka_weights": [
          1,
          1,
          1,
          1,
          1
      ],
      "n_dims_per_step": -1
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 10
• per_device_eval_batch_size: 10
• num_train_epochs: 30
• multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 10
• per_device_eval_batch_size: 10
• 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: 5e-05
• weight_decay: 0.0
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1
• num_train_epochs: 30
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.0
• 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}
• 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
• 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: batch_sampler
• multi_dataset_batch_sampler: round_robin

Training Logs

Epoch	Step	Training Loss	cosine_map@100
0.9901	200	-	0.6360
1.0	202	-	0.6399
1.9802	400	-	0.6686
2.0	404	-	0.6670
2.4752	500	2.6222	-
2.9703	600	-	0.6943
3.0	606	-	0.6864
3.9604	800	-	0.7016
4.0	808	-	0.7064
4.9505	1000	0.5981	0.7301

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 3.2.1
• Transformers: 4.44.2
• PyTorch: 2.5.0+cu121
• Accelerate: 0.34.2
• Datasets: 3.1.0
• 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}
}