yahyaabd/allstats-semantic-search-model-v1-3-origin

SentenceTransformer based on sentence-transformers/paraphrase-multilingual-mpnet-base-v2

This is a sentence-transformers model finetuned from sentence-transformers/paraphrase-multilingual-mpnet-base-v2 on the allstats-semantic-search-synthetic-dataset-v1 dataset. 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/paraphrase-multilingual-mpnet-base-v2
• Maximum Sequence Length: 128 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • allstats-semantic-search-synthetic-dataset-v1

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: XLMRobertaModel 
  (1): Pooling({'word_embedding_dimension': 768, '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-semantic-search-model-v1-3")
# Run inference
sentences = [
    'perubahan nilai tukar petani bulan mei 2017',
    'Perkembangan Nilai Tukar Petani Mei 2017',
    'Statistik Restoran/Rumah Makan Tahun 2014',
]
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

Semantic Similarity

• Datasets: allstats-semantic-search-v1-3-dev and allstat-semantic-search-v1-3-test
• Evaluated with EmbeddingSimilarityEvaluator

Metric	allstats-semantic-search-v1-3-dev	allstat-semantic-search-v1-3-test
pearson_cosine	0.9956	0.9955
spearman_cosine	0.9588	0.9583

Training Details

Training Dataset

allstats-semantic-search-synthetic-dataset-v1

• Dataset: allstats-semantic-search-synthetic-dataset-v1 at b13c0a7
• Size: 212,940 training samples
• Columns: query, doc, and label
• Approximate statistics based on the first 1000 samples:

  	query	doc	label
  type	string	string	float
  details	min: 5 tokens mean: 11.46 tokens max: 34 tokens	min: 5 tokens mean: 14.47 tokens max: 54 tokens	min: 0.0 mean: 0.5 max: 1.05

• Samples:

  query	doc	label
  aDta industri besar dan sedang Indonesia 2008	Statistik Industri Besar dan Sedang Indonesia 2008	0.9
  profil bisnis konstruksi individu jawa barat 2022	Statistik Industri Manufaktur Indonesia 2015 - Bahan Baku	0.15
  data statistik ekonomi indonesia	Nilai Tukar Valuta Asing di Indonesia 2014	0.08

• Loss: CosineSimilarityLoss with these parameters:

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

Evaluation Dataset

allstats-semantic-search-synthetic-dataset-v1

• Dataset: allstats-semantic-search-synthetic-dataset-v1 at b13c0a7
• Size: 26,618 evaluation samples
• Columns: query, doc, and label
• Approximate statistics based on the first 1000 samples:

  	query	doc	label
  type	string	string	float
  details	min: 5 tokens mean: 11.38 tokens max: 34 tokens	min: 4 tokens mean: 14.63 tokens max: 55 tokens	min: 0.0 mean: 0.51 max: 1.0

• Samples:

  query	doc	label
  tahun berapa ekspor naik 2,37% dan impor naik 30,30%?	Bulan November 2006 Ekspor Naik 2,37 % dan Impor Naik 30,30 %	1.0
  Berapa produksi padi pada tahun 2023?	Produksi padi tahun lainnya	0.0
  data statistik solus per aqua 2015	Statistik Solus Per Aqua (SPA) 2015	0.97

• 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: 32
• per_device_eval_batch_size: 32
• num_train_epochs: 12
• warmup_ratio: 0.1
• fp16: True

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: 0.0
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1.0
• num_train_epochs: 12
• 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: 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: 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: 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: 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
• 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
• use_liger_kernel: False
• eval_use_gather_object: False
• average_tokens_across_devices: False
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional

Training Logs

Click to expand

Epoch	Step	Training Loss	Validation Loss	allstats-semantic-search-v1-3-dev_spearman_cosine	allstat-semantic-search-v1-3-test_spearman_cosine
0.0751	500	0.0653	0.0400	0.7035	-
0.1503	1000	0.0361	0.0296	0.7310	-
0.2254	1500	0.0278	0.0226	0.7669	-
0.3005	2000	0.0226	0.0195	0.7748	-
0.3757	2500	0.0208	0.0183	0.7769	-
0.4508	3000	0.0184	0.0172	0.7994	-
0.5259	3500	0.0179	0.0159	0.7931	-
0.6011	4000	0.0159	0.0155	0.7966	-
0.6762	4500	0.0161	0.0150	0.8047	-
0.7513	5000	0.0163	0.0153	0.7910	-
0.8264	5500	0.0158	0.0155	0.7956	-
0.9016	6000	0.0149	0.0141	0.8148	-
0.9767	6500	0.0149	0.0145	0.8287	-
1.0518	7000	0.0148	0.0150	0.7933	-
1.1270	7500	0.0131	0.0136	0.8083	-
1.2021	8000	0.0124	0.0131	0.8173	-
1.2772	8500	0.0133	0.0130	0.8117	-
1.3524	9000	0.012	0.0126	0.8259	-
1.4275	9500	0.0119	0.0120	0.8178	-
1.5026	10000	0.0116	0.0118	0.8332	-
1.5778	10500	0.0132	0.0123	0.8108	-
1.6529	11000	0.0114	0.0111	0.8365	-
1.7280	11500	0.0105	0.0109	0.8235	-
1.8032	12000	0.0107	0.0105	0.8445	-
1.8783	12500	0.0106	0.0101	0.8330	-
1.9534	13000	0.0095	0.0096	0.8437	-
2.0285	13500	0.0093	0.0094	0.8417	-
2.1037	14000	0.0079	0.0093	0.8485	-
2.1788	14500	0.008	0.0089	0.8422	-
2.2539	15000	0.0081	0.0086	0.8485	-
2.3291	15500	0.008	0.0084	0.8530	-
2.4042	16000	0.007	0.0084	0.8597	-
2.4793	16500	0.0081	0.0087	0.8499	-
2.5545	17000	0.0078	0.0078	0.8577	-
2.6296	17500	0.007	0.0080	0.8559	-
2.7047	18000	0.0072	0.0078	0.8569	-
2.7799	18500	0.0069	0.0079	0.8579	-
2.8550	19000	0.0064	0.0072	0.8693	-
2.9301	19500	0.0064	0.0070	0.8747	-
3.0053	20000	0.0061	0.0068	0.8757	-
3.0804	20500	0.0052	0.0069	0.8727	-
3.1555	21000	0.005	0.0067	0.8734	-
3.2307	21500	0.0054	0.0065	0.8727	-
3.3058	22000	0.0058	0.0070	0.8715	-
3.3809	22500	0.0056	0.0066	0.8724	-
3.4560	23000	0.0056	0.0070	0.8740	-
3.5312	23500	0.0054	0.0060	0.8775	-
3.6063	24000	0.0051	0.0062	0.8746	-
3.6814	24500	0.0047	0.0060	0.8765	-
3.7566	25000	0.0051	0.0067	0.8783	-
3.8317	25500	0.0048	0.0058	0.8824	-
3.9068	26000	0.0048	0.0059	0.8862	-
3.9820	26500	0.005	0.0056	0.8853	-
4.0571	27000	0.0042	0.0053	0.8868	-
4.1322	27500	0.0036	0.0056	0.8893	-
4.2074	28000	0.0041	0.0052	0.8954	-
4.2825	28500	0.0041	0.0050	0.8943	-
4.3576	29000	0.0036	0.0050	0.8890	-
4.4328	29500	0.0036	0.0046	0.8990	-
4.5079	30000	0.0038	0.0051	0.8934	-
4.5830	30500	0.0037	0.0049	0.9011	-
4.6582	31000	0.0036	0.0049	0.9000	-
4.7333	31500	0.0041	0.0052	0.8938	-
4.8084	32000	0.004	0.0049	0.8971	-
4.8835	32500	0.0038	0.0043	0.9023	-
4.9587	33000	0.0036	0.0044	0.9006	-
5.0338	33500	0.0032	0.0043	0.9042	-
5.1089	34000	0.0031	0.0042	0.9054	-
5.1841	34500	0.0028	0.0042	0.9052	-
5.2592	35000	0.0028	0.0043	0.9065	-
5.3343	35500	0.003	0.0041	0.9093	-
5.4095	36000	0.0029	0.0042	0.9084	-
5.4846	36500	0.0029	0.0044	0.9078	-
5.5597	37000	0.0027	0.0043	0.9062	-
5.6349	37500	0.003	0.0039	0.9101	-
5.7100	38000	0.0027	0.0041	0.9092	-
5.7851	38500	0.0025	0.0039	0.9140	-
5.8603	39000	0.0027	0.0037	0.9138	-
5.9354	39500	0.0027	0.0037	0.9137	-
6.0105	40000	0.0027	0.0036	0.9162	-
6.0856	40500	0.002	0.0035	0.9209	-
6.1608	41000	0.0021	0.0037	0.9180	-
6.2359	41500	0.0023	0.0036	0.9183	-
6.3110	42000	0.0024	0.0035	0.9218	-
6.3862	42500	0.002	0.0033	0.9216	-
6.4613	43000	0.0024	0.0035	0.9220	-
6.5364	43500	0.0018	0.0034	0.9232	-
6.6116	44000	0.0021	0.0033	0.9236	-
6.6867	44500	0.0021	0.0035	0.9225	-
6.7618	45000	0.0027	0.0031	0.9227	-
6.8370	45500	0.0019	0.0032	0.9242	-
6.9121	46000	0.0022	0.0033	0.9224	-
6.9872	46500	0.0022	0.0030	0.9252	-
7.0624	47000	0.0017	0.0029	0.9294	-
7.1375	47500	0.0014	0.0028	0.9304	-
7.2126	48000	0.0015	0.0028	0.9324	-
7.2878	48500	0.0014	0.0030	0.9313	-
7.3629	49000	0.0015	0.0029	0.9333	-
7.4380	49500	0.0015	0.0028	0.9342	-
7.5131	50000	0.0018	0.0030	0.9261	-
7.5883	50500	0.0016	0.0030	0.9329	-
7.6634	51000	0.0019	0.0026	0.9334	-
7.7385	51500	0.0018	0.0029	0.9336	-
7.8137	52000	0.0016	0.0026	0.9353	-
7.8888	52500	0.0016	0.0026	0.9351	-
7.9639	53000	0.0017	0.0024	0.9356	-
8.0391	53500	0.0013	0.0023	0.9396	-
8.1142	54000	0.0012	0.0024	0.9390	-
8.1893	54500	0.001	0.0024	0.9421	-
8.2645	55000	0.0012	0.0024	0.9406	-
8.3396	55500	0.0012	0.0023	0.9407	-
8.4147	56000	0.0012	0.0024	0.9398	-
8.4899	56500	0.0012	0.0024	0.9412	-
8.5650	57000	0.0014	0.0024	0.9397	-
8.6401	57500	0.0013	0.0023	0.9411	-
8.7153	58000	0.0013	0.0023	0.9418	-
8.7904	58500	0.0014	0.0022	0.9432	-
8.8655	59000	0.0011	0.0022	0.9448	-
8.9406	59500	0.0012	0.0022	0.9455	-
9.0158	60000	0.0012	0.0021	0.9453	-
9.0909	60500	0.0009	0.0021	0.9461	-
9.1660	61000	0.0009	0.0021	0.9465	-
9.2412	61500	0.0009	0.0021	0.9471	-
9.3163	62000	0.0009	0.0021	0.9477	-
9.3914	62500	0.0008	0.0020	0.9482	-
9.4666	63000	0.0012	0.0020	0.9478	-
9.5417	63500	0.0009	0.0020	0.9479	-
9.6168	64000	0.0009	0.0020	0.9485	-
9.6920	64500	0.0011	0.0020	0.9492	-
9.7671	65000	0.0008	0.0019	0.9497	-
9.8422	65500	0.001	0.0019	0.9504	-
9.9174	66000	0.0009	0.0019	0.9518	-
9.9925	66500	0.0009	0.0019	0.9510	-
10.0676	67000	0.0008	0.0018	0.9517	-
10.1427	67500	0.0007	0.0018	0.9524	-
10.2179	68000	0.0007	0.0018	0.9521	-
10.2930	68500	0.0008	0.0019	0.9526	-
10.3681	69000	0.0007	0.0019	0.9529	-
10.4433	69500	0.0008	0.0018	0.9541	-
10.5184	70000	0.0007	0.0017	0.9551	-
10.5935	70500	0.0007	0.0018	0.9550	-
10.6687	71000	0.0008	0.0017	0.9554	-
10.7438	71500	0.0007	0.0017	0.9558	-
10.8189	72000	0.0007	0.0018	0.9558	-
10.8941	72500	0.0007	0.0018	0.9562	-
10.9692	73000	0.0009	0.0017	0.9559	-
11.0443	73500	0.0005	0.0017	0.9571	-
11.1195	74000	0.0006	0.0017	0.9570	-
11.1946	74500	0.0005	0.0017	0.9573	-
11.2697	75000	0.0005	0.0017	0.9574	-
11.3449	75500	0.0006	0.0017	0.9576	-
11.4200	76000	0.0006	0.0017	0.9577	-
11.4951	76500	0.0006	0.0017	0.9577	-
11.5702	77000	0.0005	0.0016	0.9582	-
11.6454	77500	0.0006	0.0017	0.9583	-
11.7205	78000	0.0005	0.0016	0.9584	-
11.7956	78500	0.0005	0.0016	0.9587	-
11.8708	79000	0.0005	0.0016	0.9588	-
11.9459	79500	0.0005	0.0016	0.9588	-
12.0	79860	-	-	-	0.9583

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 3.3.1
• Transformers: 4.47.1
• PyTorch: 2.2.2+cu121
• Accelerate: 1.2.1
• 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",
}