yahyaabd / allstats-search-mini-v2

SentenceTransformer based on yahyaabd/allstats-search-mini-v1-1-mnrl

This is a sentence-transformers model finetuned from yahyaabd/allstats-search-mini-v1-1-mnrl on the bps-pub-cosine-pairs dataset. 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: yahyaabd/allstats-search-mini-v1-1-mnrl
• Maximum Sequence Length: 128 tokens
• Output Dimensionality: 384 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • bps-pub-cosine-pairs

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': 128, '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("yahyaabd/allstats-search-mini-v2")
# Run inference
sentences = [
    'q-786',
    'Angka Kematian Bayi oper P#rovinsi',
    'f3b02f2b6706e104ea9d5b74',
]
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

Semantic Similarity

• Datasets: sts-dev and sts-test
• Evaluated with EmbeddingSimilarityEvaluator

Training Details

Training Dataset

bps-pub-cosine-pairs

• Dataset: bps-pub-cosine-pairs at 038a9de
• Size: 64,260 training samples
• Columns: query_id , query , corpus_id , title , and score
• Approximate statistics based on the first 1000 samples:

  	query_id	query	corpus_id	title	score
  type	string	string	string	string	float
  details	min: 4 tokens mean: 5.18 tokens max: 6 tokens	min: 4 tokens mean: 13.33 tokens max: 38 tokens	min: 7 tokens mean: 17.38 tokens max: 22 tokens	min: 5 tokens mean: 13.13 tokens max: 30 tokens	min: 0.1 mean: 0.56 max: 0.9

• Samples:

  query_id	query	corpus_id	title	score
  q-1599	Nilai Tukar Nelayan	0b0da8fc2b6af9329a6d9cfe	Statistik Hotel dan Akomodasi Lainnya di Indonesia 2013	0.1
  q-1599	nilai tukar nelayan	0b0da8fc2b6af9329a6d9cfe	Statistik Hotel dan Akomodasi Lainnya di Indonesia 2013	0.1
  q-1599	NILAI TUKAR NELAYAN	0b0da8fc2b6af9329a6d9cfe	Statistik Hotel dan Akomodasi Lainnya di Indonesia 2013	0.1

• Loss: CosineSimilarityLoss with these parameters:

  {
      "loss_fct": "torch.nn.modules.loss.MSELoss"
  }
  

Evaluation Dataset

bps-pub-cosine-pairs

• Dataset: bps-pub-cosine-pairs at 038a9de
• Size: 8,067 evaluation samples
• Columns: query_id , query , corpus_id , title , and score
• Approximate statistics based on the first 1000 samples:

  	query_id	query	corpus_id	title	score
  type	string	string	string	string	float
  details	min: 4 tokens mean: 5.2 tokens max: 6 tokens	min: 4 tokens mean: 12.77 tokens max: 33 tokens	min: 13 tokens mean: 17.25 tokens max: 23 tokens	min: 5 tokens mean: 13.37 tokens max: 38 tokens	min: 0.1 mean: 0.57 max: 0.9

• Samples:

  query_id	query	corpus_id	title	score
  q-1273	Sosek Desember 2021	b7890a143bc751d1d84dcf4a	Laporan Bulanan Data Sosial Ekonomi Desember 2021	0.9
  q-1273	sosek desember 2021	b7890a143bc751d1d84dcf4a	Laporan Bulanan Data Sosial Ekonomi Desember 2021	0.9
  q-1273	SOSEK DESEMBER 2021	b7890a143bc751d1d84dcf4a	Laporan Bulanan Data Sosial Ekonomi Desember 2021	0.9

• Loss: CosineSimilarityLoss with these parameters:

  {
      "loss_fct": "torch.nn.modules.loss.MSELoss"
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy : steps
• per_device_train_batch_size : 64
• per_device_eval_batch_size : 64
• learning_rate : 1e-05
• num_train_epochs : 2
• warmup_ratio : 0.1
• fp16 : True
• load_best_model_at_end : True
• label_smoothing_factor : 0.01
• eval_on_start : True

All Hyperparameters

Training Logs

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 3.4.0
• Transformers: 4.48.1
• PyTorch: 2.5.1+cu124
• Accelerate: 1.3.0
• Datasets: 3.2.0
• Tokenizers: 0.21.0

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",
}