Add new SentenceTransformer model

da52d96 verified 7 months ago

27 kB

metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:10998
  - loss:MultipleNegativesRankingLoss
base_model: yahyaabd/bps-custom-tokenizer-paraphrase-multilingual-MiniLM-L12-v2
widget:
  - source_sentence: Laporan neraca arus dana dalam Rupiah miliar untuk Q2 2011
    sentences:
      - Ringkasan Neraca Arus Dana, Triwulan II, 2011*), (Miliar Rupiah)
      - >-
        Tingkat Inflasi Harga Konsumen Nasional Tahun Kalender (Y-to-D) 1
        (2022=100)
      - >-
        Jumlah Pendapatan Setelah Pajak Menurut Golongan Rumah Tangga (miliar
        rupiah), 2000, 2005, dan 2008
  - source_sentence: >-
      Pinjaman bank umum Rp  valas dan BPR Rp ke pihak nonbank data historis
      sampai 2022
    sentences:
      - >-
        Rata-rata Upah/Gaji Bersih Sebulan Buruh/Karyawan/Pegawai Menurut
        Pendidikan Tertinggi yang Ditamatkan dan Jenis Pekerjaan Utama, 2023
      - >-
        Posisi Kredit/Pembiayaan (Rupiah dan Valuta Asing) pada Bank Umum
        Menurut Kelompok Bank dan BPR/BPR Syariah (Rupiah) kepada Pihak Bukan
        Bank (miliar rupiah), 2014-2023
      - >-
        Rata-rata Konsumsi dan Pengeluaran Perkapita Seminggu Menurut Komoditi
        Makanan dan Golongan Pengeluaran per Kapita Seminggu di Provinsi
        Sulawesi Barat, 2018-2023
  - source_sentence: Infant mortality rate Indonesia per kabupaten/kota tahun 2020
    sentences:
      - >-
        Rata-rata Pendapatan bersih Berusaha Sendiri menurut Provinsi dan
        Pendidikan yang Ditamatkan, 2023
      - >-
        Rekapitulasi Luas Penutupan Lahan Hutan dan Non Hutan Menurut Provinsi
        Tahun 2014-2022 (Ribu Ha)
      - >-
        Angka Kematian Bayi/AKB (Infant Mortality Rate/IMR) Hasil Long Form
        SP2020 Menurut Provinsi/Kabupaten/Kota, 2020
  - source_sentence: >-
      Berapa pendapatan bulanan ratarata pengusaha di Indonesia Data per lokasi
      bedah sektor 2024
    sentences:
      - >-
        Bongkar Muat Barang Antar Pulau dan Luar Negeri di Pelabuhan Indonesia
        Tahun 1988-2022 (Ribu ton)
      - >-
        Persentase Rumah Tangga Menurut Provinsi dan Kepemilikan Kendaraan
        Bermotor, 2013-2014
      - >-
        Rata-rata Pendapatan Bersih Berusaha Sendiri Menurut Provinsi dan
        Lapangan Pekerjaan Utama, 2024
  - source_sentence: Informasi lengkap dan terbaru mengenai statistik edukasi
    sentences:
      - >-
        Struktur Ongkos Riil Usaha Ternak dan Unggas di Rumah Tangga dengan Pola
        Pemeliharaan Dikandangkan, 2017
      - >-
        Rata-rata Pendapatan Bersih Berusaha Sendiri Menurut Provinsi dan
        Kelompok Umur (ribu rupiah), 2018
      - Statistik Pendidikan Tahunan
datasets:
  - yahyaabd/statictable-triplets-all
pipeline_tag: sentence-similarity
library_name: sentence-transformers
metrics:
  - cosine_accuracy@1
  - cosine_accuracy@5
  - cosine_accuracy@10
  - cosine_precision@1
  - cosine_precision@5
  - cosine_precision@10
  - cosine_recall@1
  - cosine_recall@5
  - cosine_recall@10
  - cosine_ndcg@1
  - cosine_ndcg@5
  - cosine_ndcg@10
  - cosine_mrr@1
  - cosine_mrr@5
  - cosine_mrr@10
  - cosine_map@1
  - cosine_map@5
  - cosine_map@10
model-index:
  - name: >-
      SentenceTransformer based on
      yahyaabd/bps-custom-tokenizer-paraphrase-multilingual-MiniLM-L12-v2
    results:
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: bps statictable ir
          type: bps-statictable-ir
        metrics:
          - type: cosine_accuracy@1
            value: 0.8599348534201955
            name: Cosine Accuracy@1
          - type: cosine_accuracy@5
            value: 0.9804560260586319
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.9869706840390879
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.8599348534201955
            name: Cosine Precision@1
          - type: cosine_precision@5
            value: 0.23257328990228016
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.13908794788273618
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.6663188638274197
            name: Cosine Recall@1
          - type: cosine_recall@5
            value: 0.7919003338686821
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.8157207271760252
            name: Cosine Recall@10
          - type: cosine_ndcg@1
            value: 0.8599348534201955
            name: Cosine Ndcg@1
          - type: cosine_ndcg@5
            value: 0.8115433718814501
            name: Cosine Ndcg@5
          - type: cosine_ndcg@10
            value: 0.8115549383345654
            name: Cosine Ndcg@10
          - type: cosine_mrr@1
            value: 0.8599348534201955
            name: Cosine Mrr@1
          - type: cosine_mrr@5
            value: 0.9118892508143324
            name: Cosine Mrr@5
          - type: cosine_mrr@10
            value: 0.9127940644227288
            name: Cosine Mrr@10
          - type: cosine_map@1
            value: 0.8599348534201955
            name: Cosine Map@1
          - type: cosine_map@5
            value: 0.7617137169743033
            name: Cosine Map@5
          - type: cosine_map@10
            value: 0.7557582338038363
            name: Cosine Map@10

SentenceTransformer based on yahyaabd/bps-custom-tokenizer-paraphrase-multilingual-MiniLM-L12-v2

This is a sentence-transformers model finetuned from yahyaabd/bps-custom-tokenizer-paraphrase-multilingual-MiniLM-L12-v2 on the statictable-triplets-all 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/bps-custom-tokenizer-paraphrase-multilingual-MiniLM-L12-v2
Maximum Sequence Length: 512 tokens
Output Dimensionality: 384 dimensions
Similarity Function: Cosine Similarity
Training Dataset:
- statictable-triplets-all

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("yahyaabd/bps-custom-tokenizer-paraphrase-multilingual-MiniLM-L12-v2-mnrl-2")
# Run inference
sentences = [
    'Informasi lengkap dan terbaru mengenai statistik edukasi',
    'Statistik Pendidikan Tahunan',
    'Struktur Ongkos Riil Usaha Ternak dan Unggas di Rumah Tangga dengan Pola Pemeliharaan Dikandangkan, 2017',
]
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

Information Retrieval

Dataset: bps-statictable-ir
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.8599
cosine_accuracy@5	0.9805
cosine_accuracy@10	0.987
cosine_precision@1	0.8599
cosine_precision@5	0.2326
cosine_precision@10	0.1391
cosine_recall@1	0.6663
cosine_recall@5	0.7919
cosine_recall@10	0.8157
cosine_ndcg@1	0.8599
cosine_ndcg@5	0.8115
cosine_ndcg@10	0.8116
cosine_mrr@1	0.8599
cosine_mrr@5	0.9119
cosine_mrr@10	0.9128
cosine_map@1	0.8599
cosine_map@5	0.7617
cosine_map@10	0.7558

Training Details

Training Dataset

statictable-triplets-all

Dataset: statictable-triplets-all at 0ef226c
Size: 10,998 training samples
Columns: query, positive, and negative

Approximate statistics based on the first 1000 samples:

	query	positive	negative
type	string	string	string
details	min: 6 tokens mean: 17.25 tokens max: 33 tokens	min: 5 tokens mean: 25.66 tokens max: 58 tokens	min: 4 tokens mean: 25.64 tokens max: 58 tokens

Samples:

query	positive	negative
`Neraca arus kas triwulan II 2005 (ringkasan, )`	`Ringkasan Neraca Arus Dana, Triwulan Kedua, 2005, (Miliar Rupiah)`	`Rata-rata Upah/Gaji Bersih Sebulan Buruh/Karyawan/Pegawai Menurut Provinsi dan Jenis Pekerjaan Utama (Rupiah), 2017`
`Hasil tangkapan ikan per provinsi, bedakan jenis penangkapan, 2013`	`Produksi Perikanan Tangkap Menurut Provinsi dan Jenis Penangkapan, 2000-2020`	`Ringkasan Neraca Arus Dana, Triwulan II, 2006, (Miliar Rupiah)`
`Bagaimana perubahan distribusi pengeluaran?`	`Persentase Perkembangan Distribusi Pengeluaran`	`Angka Kematian Bayi/AKB (Infant Mortality Rate/IMR) Menurut Provinsi, 1971-2020`

Loss: MultipleNegativesRankingLoss with these parameters:

{
    "scale": 20.0,
    "similarity_fct": "cos_sim"
}

Evaluation Dataset

statictable-triplets-all

Dataset: statictable-triplets-all at 0ef226c
Size: 10,998 evaluation samples
Columns: query, positive, and negative

Approximate statistics based on the first 1000 samples:

	query	positive	negative
type	string	string	string
details	min: 4 tokens mean: 17.25 tokens max: 34 tokens	min: 4 tokens mean: 25.44 tokens max: 58 tokens	min: 5 tokens mean: 25.23 tokens max: 58 tokens

Samples:

query	positive	negative
`Data total penghasilan berbagai golongan rumah tangga setelah dipotong pajak, tahun 2000 (dalam )`	`Jumlah Pendapatan Setelah Pajak Menurut Golongan Rumah Tangga (miliar rupiah), 2000, 2005, dan 2008`	`Indeks Harga Konsumen per Kelompok di 82 Kota 1 (2012=100)`
`Bagaimana perkembangan impor barang modal pada tahun 2020`	`Impor Barang Modal, 1996-2023`	`Indeks Harga yang Diterima Petani (It), Indes Harga yang Dibayar Petani (Ib), dan Nilai Tukar Petani Subsektor Hortikultura (NTPH) di Indonesia (2007=100), 2008-2016`
`Konsumsi makanan per orang di Kalut: data mingguan, beda kelompok pengeluaran (2018)`	`Rata-rata Konsumsi dan Pengeluaran Perkapita Seminggu Menurut Komoditi Makanan dan Golongan Pengeluaran per Kapita Seminggu di Provinsi Kalimantan Utara, 2018-2023`	`Ekspor Kimia Dasar Organik yang Bersumber dari Hasil Pertanian menurut Negara Tujuan Utama, 2012-2023`

Loss: MultipleNegativesRankingLoss with these parameters:

{
    "scale": 20.0,
    "similarity_fct": "cos_sim"
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 32
per_device_eval_batch_size: 32
weight_decay: 0.01
warmup_ratio: 0.1
save_on_each_node: True
fp16: True
dataloader_num_workers: 2
load_best_model_at_end: True
eval_on_start: True
batch_sampler: no_duplicates

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: steps
prediction_loss_only: True
per_device_train_batch_size: 32
per_device_eval_batch_size: 32
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: 0.01
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: linear
lr_scheduler_kwargs: {}
warmup_ratio: 0.1
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: True
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: True
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: 2
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}
tp_size: 0
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: None
hub_always_push: False
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
include_inputs_for_metrics: False
include_for_metrics: []
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
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: True
use_liger_kernel: False
eval_use_gather_object: False
average_tokens_across_devices: False
prompts: None
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	Validation Loss	bps-statictable-ir_cosine_ndcg@10
0	0	-	2.2208	0.3476
0.0645	20	1.7839	0.9899	0.5271
0.1290	40	0.7326	0.4401	0.7019
0.1935	60	0.3811	0.2612	0.7584
0.2581	80	0.2068	0.2111	0.7612
0.3226	100	0.2206	0.1526	0.7748
0.3871	120	0.1547	0.1065	0.7934
0.4516	140	0.1196	0.0895	0.7880
0.5161	160	0.1107	0.0821	0.8045
0.5806	180	0.1253	0.0737	0.7828
0.6452	200	0.0915	0.0636	0.8081
0.7097	220	0.0592	0.0555	0.8140
0.7742	240	0.055	0.0535	0.7992
0.8387	260	0.0531	0.0487	0.8005
0.9032	280	0.0626	0.0429	0.8035
0.9677	300	0.0406	0.0407	0.8033
1.0323	320	0.034	0.0430	0.8058
1.0968	340	0.0327	0.0392	0.8070
1.1613	360	0.0385	0.0425	0.8006
1.2258	380	0.0233	0.0347	0.8053
1.2903	400	0.027	0.0339	0.8111
1.3548	420	0.0323	0.0300	0.8046
1.4194	440	0.0308	0.0262	0.8126
1.4839	460	0.0343	0.0277	0.7961
1.5484	480	0.0192	0.0232	0.8080
1.6129	500	0.0248	0.0248	0.8057
1.6774	520	0.0178	0.0250	0.8062
1.7419	540	0.0158	0.0228	0.8096
1.8065	560	0.0171	0.0233	0.8073
1.8710	580	0.0204	0.0218	0.8178
1.9355	600	0.0261	0.0214	0.8204
2.0	620	0.0132	0.0215	0.8166
2.0645	640	0.0174	0.0189	0.8169
2.1290	660	0.0095	0.0185	0.8202
2.1935	680	0.0186	0.0173	0.8173
2.2581	700	0.0241	0.0168	0.8174
2.3226	720	0.0152	0.0158	0.8163
2.3871	740	0.0197	0.0158	0.8128
2.4516	760	0.0119	0.0156	0.8122
2.5161	780	0.0128	0.0151	0.8118
2.5806	800	0.0162	0.0148	0.8114
2.6452	820	0.011	0.0143	0.8117
2.7097	840	0.0098	0.0138	0.8128
2.7742	860	0.0092	0.0135	0.8111
2.8387	880	0.0102	0.0127	0.8109
2.9032	900	0.0118	0.0126	0.8115
2.9677	920	0.0128	0.0126	0.8116

The bold row denotes the saved checkpoint.

Framework Versions

Python: 3.11.12
Sentence Transformers: 3.4.0
Transformers: 4.51.3
PyTorch: 2.6.0+cu124
Accelerate: 1.6.0
Datasets: 3.2.0
Tokenizers: 0.21.1

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",
}

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}
}