splendor1811/BGE-base-banking-ONE

BGE base banking-domain

This is a sentence-transformers model finetuned from BAAI/bge-m3 on the json dataset. It maps sentences & paragraphs to a 1024-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: BAAI/bge-m3
• Maximum Sequence Length: 8192 tokens
• Output Dimensionality: 1024 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • json
• Language: vi
• 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': 8192, 'do_lower_case': False}) with Transformer model: XLMRobertaModel 
  (1): Pooling({'word_embedding_dimension': 1024, '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})
  (2): Normalize()
)

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("splendor1811/BGE-base-banking-ONE")
# Run inference
sentences = [
    'Các giao dịch sử dụng thẻ tín dụng của tôi ',
    'Hướng dẫn xem lịch sử ',
    'Thiết bị loa sẽ được bảo hành trong 12 tháng. Nếu có vấn đề về sản phẩm trong quá trình sử dụng, Bạn vui lòng đến Trung Tâm Bảo Hành Phong Vũ gần nhất hoặc liên hệ hotline: 1800 6865 để được hỗ trợ bảo hành.\nThông tin về cửa hàng bảo hành Phong Vũ như sau:\n+ Miền Bắc: Tầng 3, số 62 Trần Đại Nghĩa, Phường Đồng Tâm, Quận Hai Bà Trưng, TP. Hà Nội.\n+ Miền Nam: 132E Cách Mạng Tháng 8, Phường 9, Quận 3, TP. Hồ Chí Minh.\n+ Miền Trung: Tầng 2, 14-16-18 Nguyễn Văn Linh, Phường Nam Dương, Quận Hải Châu, TP. Đà Nẵng.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]

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

Evaluation

Metrics

Information Retrieval

• Dataset: dim_1024
• Evaluated with InformationRetrievalEvaluator with these parameters:

  {
      "truncate_dim": 1024
  }
  

Metric	Value
cosine_accuracy@1	0.6794
cosine_accuracy@3	0.6794
cosine_accuracy@5	0.6794
cosine_accuracy@10	0.7495
cosine_precision@1	0.6794
cosine_precision@3	0.6794
cosine_precision@5	0.6794
cosine_precision@10	0.6476
cosine_recall@1	0.0645
cosine_recall@3	0.1935
cosine_recall@5	0.3226
cosine_recall@10	0.6094
cosine_ndcg@10	0.6841
cosine_mrr@10	0.6864
cosine_map@100	0.7415

Information Retrieval

• Dataset: dim_768
• Evaluated with InformationRetrievalEvaluator with these parameters:

  {
      "truncate_dim": 768
  }
  

Metric	Value
cosine_accuracy@1	0.6833
cosine_accuracy@3	0.6833
cosine_accuracy@5	0.6833
cosine_accuracy@10	0.7502
cosine_precision@1	0.6833
cosine_precision@3	0.6833
cosine_precision@5	0.6833
cosine_precision@10	0.6506
cosine_recall@1	0.065
cosine_recall@3	0.1951
cosine_recall@5	0.3252
cosine_recall@10	0.6137
cosine_ndcg@10	0.6878
cosine_mrr@10	0.69
cosine_map@100	0.7446

Information Retrieval

• Dataset: dim_512
• Evaluated with InformationRetrievalEvaluator with these parameters:

  {
      "truncate_dim": 512
  }
  

Metric	Value
cosine_accuracy@1	0.6742
cosine_accuracy@3	0.6742
cosine_accuracy@5	0.6742
cosine_accuracy@10	0.745
cosine_precision@1	0.6742
cosine_precision@3	0.6742
cosine_precision@5	0.6742
cosine_precision@10	0.6426
cosine_recall@1	0.064
cosine_recall@3	0.1919
cosine_recall@5	0.3198
cosine_recall@10	0.6041
cosine_ndcg@10	0.679
cosine_mrr@10	0.6813
cosine_map@100	0.7379

Information Retrieval

• Dataset: dim_256
• Evaluated with InformationRetrievalEvaluator with these parameters:

  {
      "truncate_dim": 256
  }
  

Metric	Value
cosine_accuracy@1	0.6684
cosine_accuracy@3	0.6684
cosine_accuracy@5	0.6684
cosine_accuracy@10	0.7346
cosine_precision@1	0.6684
cosine_precision@3	0.6684
cosine_precision@5	0.6684
cosine_precision@10	0.6369
cosine_recall@1	0.0631
cosine_recall@3	0.1892
cosine_recall@5	0.3153
cosine_recall@10	0.5953
cosine_ndcg@10	0.6728
cosine_mrr@10	0.675
cosine_map@100	0.7314

Information Retrieval

• Dataset: dim_128
• Evaluated with InformationRetrievalEvaluator with these parameters:

  {
      "truncate_dim": 128
  }
  

Metric	Value
cosine_accuracy@1	0.636
cosine_accuracy@3	0.636
cosine_accuracy@5	0.636
cosine_accuracy@10	0.7099
cosine_precision@1	0.636
cosine_precision@3	0.636
cosine_precision@5	0.636
cosine_precision@10	0.6071
cosine_recall@1	0.0601
cosine_recall@3	0.1803
cosine_recall@5	0.3005
cosine_recall@10	0.5685
cosine_ndcg@10	0.6409
cosine_mrr@10	0.6433
cosine_map@100	0.7035

Training Details

Training Dataset

json

• Dataset: json
• Size: 13,863 training samples
• Columns: anchor and positive
• Approximate statistics based on the first 1000 samples:

  	anchor	positive
  type	string	string
  details	min: 3 tokens mean: 22.35 tokens max: 61 tokens	min: 3 tokens mean: 225.69 tokens max: 419 tokens

• Loss: MatryoshkaLoss with these parameters:

  {
      "loss": "MultipleNegativesRankingLoss",
      "matryoshka_dims": [
          1024,
          768,
          512,
          256,
          128
      ],
      "matryoshka_weights": [
          1,
          1,
          1,
          1,
          1
      ],
      "n_dims_per_step": -1
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: epoch
• per_device_train_batch_size: 32
• per_device_eval_batch_size: 16
• gradient_accumulation_steps: 16
• learning_rate: 2e-05
• num_train_epochs: 6
• lr_scheduler_type: cosine
• warmup_ratio: 0.1
• bf16: True
• tf32: True
• load_best_model_at_end: True
• optim: adamw_torch_fused
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: epoch
• prediction_loss_only: True
• per_device_train_batch_size: 32
• per_device_eval_batch_size: 16
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 16
• eval_accumulation_steps: None
• torch_empty_cache_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: 6
• max_steps: -1
• lr_scheduler_type: cosine
• 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: True
• fp16: False
• fp16_opt_level: O1
• half_precision_backend: auto
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: True
• 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: 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}
• 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_fused
• 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: False
• 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	dim_1024_cosine_ndcg@10	dim_768_cosine_ndcg@10	dim_512_cosine_ndcg@10	dim_256_cosine_ndcg@10	dim_128_cosine_ndcg@10
0.3687	10	48.8875	-	-	-	-	-
0.7373	20	23.8518	-	-	-	-	-
1.0	28	-	0.6421	0.6376	0.6334	0.6215	0.5950
1.0737	30	16.242	-	-	-	-	-
1.4424	40	13.0298	-	-	-	-	-
1.8111	50	12.8472	-	-	-	-	-
2.0	56	-	0.6764	0.6663	0.6589	0.6487	0.6127
2.1475	60	9.3195	-	-	-	-	-
2.5161	70	9.0553	-	-	-	-	-
2.8848	80	9.8082	-	-	-	-	-
3.0	84	-	0.6801	0.6792	0.6749	0.6679	0.6279
3.2212	90	7.864	-	-	-	-	-
3.5899	100	7.6955	-	-	-	-	-
3.9585	110	8.0813	-	-	-	-	-
4.0	112	-	0.6879	0.6888	0.6779	0.6645	0.6361
4.2949	120	6.899	-	-	-	-	-
4.6636	130	7.1247	-	-	-	-	-
5.0	140	6.2173	0.6841	0.6859	0.6770	0.6702	0.6410
5.3687	150	6.741	-	-	-	-	-
5.7373	160	6.5777	-	-	-	-	-
6.0	168	-	0.6841	0.6878	0.6790	0.6728	0.6409

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.11.12
• Sentence Transformers: 4.1.0
• Transformers: 4.52.4
• PyTorch: 2.6.0+cu124
• Accelerate: 1.7.0
• Datasets: 3.6.0
• 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",
}

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}
}