ivanleomk/finetuned-bge-base-en

SentenceTransformer based on BAAI/bge-base-en

This is a sentence-transformers model finetuned from BAAI/bge-base-en. 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: BAAI/bge-base-en
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 tokens
• Similarity Function: Cosine Similarity

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': True}) with Transformer model: BertModel 
  (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})
  (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("ivanleomk/finetuned-bge-base-en")
# Run inference
sentences = [
    '\nName : Viacom Solutions\nCategory: Telecom Hardware, Network Architecture\nDepartment: Engineering\nLocation: Tokyo, Japan\nAmount: 1450.67\nCard: Global Network Optimization Project\nTrip Name: unknown\n',
    '\nName : Pardalis Digital\nCategory: Data Analytics Platform, Professional Networking Service\nDepartment: Sales\nLocation: Dublin, Ireland\nAmount: 1456.75\nCard: Sales Intelligence & Networking Platform\nTrip Name: unknown\n',
    "\nName : Il Vino e L'Arte\nCategory: Culinary Experience, Cultural Event Venue\nDepartment: Marketing\nLocation: Rome, Italy\nAmount: 748.32\nCard: Cultural Engagement Dinner\nTrip Name: unknown\n",
]
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

Triplet

• Dataset: bge-base-en-train
• Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.8462
dot_accuracy	0.1538
manhattan_accuracy	0.8558
euclidean_accuracy	0.8462
max_accuracy	0.8558

Triplet

• Dataset: bge-base-en-eval
• Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.9545
dot_accuracy	0.0455
manhattan_accuracy	0.9545
euclidean_accuracy	0.9545
max_accuracy	0.9545

Training Details

Training Dataset

Unnamed Dataset

• Size: 208 training samples
• Columns: sentence and label
• Approximate statistics based on the first 208 samples:

  	sentence	label
  type	string	int
  details	min: 33 tokens mean: 39.66 tokens max: 48 tokens	0: ~4.81% 1: ~5.29% 2: ~6.25% 3: ~2.40% 4: ~3.85% 5: ~4.33% 6: ~3.85% 7: ~2.40% 8: ~4.81% 9: ~3.37% 10: ~3.85% 11: ~3.85% 12: ~4.81% 13: ~4.81% 14: ~5.29% 15: ~3.37% 16: ~4.81% 17: ~4.33% 18: ~3.85% 19: ~1.92% 20: ~2.88% 21: ~2.88% 22: ~3.37% 23: ~0.96% 24: ~4.33% 25: ~2.40% 26: ~0.96%

• Samples:

  sentence	label
  Name : Global Insights Group Category: Subscriptions & Memberships, Data Services & Analytics Department: Marketing Location: London, UK Amount: 1245.67 Card: Marketing Intelligence Fund Trip Name: unknown	0
  Name : CyberGuard Provisions Category: Security Software Solutions, Data Protection Services Department: Information Security Location: San Francisco, CA Amount: 879.92 Card: Digital Fortress Action Plan Trip Name: unknown	1
  Name : Apex Innovations Group Category: Business Consulting, Training Services Department: Executive Location: Sydney, Australia Amount: 1575.34 Card: Leadership Development Program Trip Name: unknown	2

• Loss: BatchSemiHardTripletLoss

Evaluation Dataset

Unnamed Dataset

• Size: 52 evaluation samples
• Columns: sentence and label
• Approximate statistics based on the first 52 samples:

  	sentence	label
  type	string	int
  details	min: 32 tokens mean: 40.13 tokens max: 49 tokens	0: ~5.77% 1: ~1.92% 2: ~3.85% 3: ~1.92% 4: ~1.92% 5: ~1.92% 6: ~5.77% 8: ~3.85% 9: ~7.69% 10: ~5.77% 12: ~3.85% 13: ~5.77% 14: ~3.85% 15: ~1.92% 16: ~9.62% 17: ~1.92% 18: ~1.92% 19: ~3.85% 20: ~1.92% 21: ~3.85% 22: ~5.77% 23: ~3.85% 24: ~5.77% 25: ~5.77%

• Samples:

  sentence	label
  Name : Viacom Solutions Category: Telecom Hardware, Network Architecture Department: Engineering Location: Tokyo, Japan Amount: 1450.67 Card: Global Network Optimization Project Trip Name: unknown	9
  Name : Vista Cascades Resort Category: Hospitality, Event Hosting Department: Sales Location: Orlando, FL Amount: 1823.45 Card: Annual Sales Retreat Trip Name: Q3 Strategy Session	12
  Name : ActiveHealth CoLab Category: Health Services, Wellness Solutions Department: HR Location: Amsterdam, Netherlands Amount: 745.32 Card: Corporate Wellness Partnership Trip Name: unknown	23

• Loss: BatchSemiHardTripletLoss

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 16
• learning_rate: 2e-05
• num_train_epochs: 5
• warmup_ratio: 0.1
• fp16: 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: 16
• per_device_eval_batch_size: 16
• 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: 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: 5
• 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: 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
• use_liger_kernel: False
• eval_use_gather_object: False
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	bge-base-en-eval_max_accuracy	bge-base-en-train_max_accuracy
0	0	-	0.8558
5.0	65	0.9545	-

Framework Versions

• Python: 3.11.10
• Sentence Transformers: 3.1.1
• Transformers: 4.45.2
• PyTorch: 2.5.1+cu124
• Accelerate: 1.1.1
• Datasets: 3.1.0
• Tokenizers: 0.20.3

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",
}

BatchSemiHardTripletLoss

@misc{hermans2017defense,
    title={In Defense of the Triplet Loss for Person Re-Identification},
    author={Alexander Hermans and Lucas Beyer and Bastian Leibe},
    year={2017},
    eprint={1703.07737},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}