metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:208
  - loss:BatchSemiHardTripletLoss
base_model: BAAI/bge-base-en
widget:
  - source_sentence: |

      Name : SkillAdvance Academy
      Category: Online Learning Platform, Professional Development
      Department: Engineering
      Location: Austin, TX
      Amount: 1875.67
      Card: Continuous Improvement Initiative
      Trip Name: unknown
    sentences:
      - |

        Name : Black Wolf
        Category: Luxury Vehicle Rentals, Corporate Services
        Department: Executive
        Location: Tokyo, Japan
        Amount: 1478.67
        Card: Execute Account
        Trip Name: Tokyo Summit 2023
      - |

        Name : Kreutz & Partners
        Category: Strategic Consulting
        Department: Marketing
        Location: Zurich, Switzerland
        Amount: 982.75
        Card: Digital Growth Strategy
        Trip Name: unknown
      - |

        Name : Nordiska Hosting Collective
        Category: Cloud Storage Solutions, Data Security Services
        Department: IT Operations
        Location: Helsinki, Finland
        Amount: 1439.57
        Card: Annual Data Management Plan
        Trip Name: unknown
  - source_sentence: |

      Name : FusionLink
      Category: Event Management Solutions, Digital Strategy Services
      Department: Sales
      Location: New York, NY
      Amount: 982.75
      Card: Product Launch Activation
      Trip Name: unknown
    sentences:
      - |

        Name : Globetrotter Partners
        Category: Lodging Services, Corporate Retreat Planning
        Department: Executive
        Location: Banff, Canada
        Amount: 1559.75
        Card: Leadership Development Seminar
        Trip Name: unknown
      - |

        Name : SkyHigh Consultancies
        Category: Consulting Services, Business Travel Agencies
        Department: Sales
        Location: Geneva, Switzerland
        Amount: 1349.58
        Card: Strategic Client Meetings
        Trip Name: Global Expansion Initiative
      - |

        Name : Willink Labs
        Category: Consulting Services, Professional Services
        Department: Engineering
        Location: San Francisco, CA
        Amount: 4500.0
        Card: Backend Systems Upgrade Analysis
        Trip Name: unknown
  - source_sentence: |

      Name : RBC
      Category: Transaction Processing, Financial Services
      Department: Finance
      Location: Limassol, Cyprus
      Amount: 843.56
      Card: Quarterly Financial Management
      Trip Name: unknown
    sentences:
      - |

        Name : Kepler Dynamics
        Category: Strategic Consultancy, Tech Solutions
        Department: Finance
        Location: Zurich, Switzerland
        Amount: 2375.88
        Card: Integration Strategy Review
        Trip Name: unknown
      - |

        Name : Global Interconnectivity Corp
        Category: Data Management Services, Network Infrastructure Consultants
        Department: Engineering
        Location: Zurich, Switzerland
        Amount: 1987.54
        Card: Unified Communication Rollout
        Trip Name: unknown
      - |

        Name : TechSupply Inc.
        Category: Electronics Retail, Supply Chain
        Department: Research & Development
        Location: Berlin, Germany
        Amount: 742.45
        Card: New Prototype Equipment
        Trip Name: unknown
  - source_sentence: |

      Name : EcoClean Systems
      Category: Environmental Services, Industrial Equipment Care
      Department: Office Administration
      Location: San Francisco, CA
      Amount: 952.63
      Card: Essential Facility Sustainability
      Trip Name: unknown
    sentences:
      - |

        Name : Wunder
        Category: Advanced Electronics
        Department: Operations
        Location: Munich, Germany
        Amount: 1643.87
        Card: Enterprise Systems Initiative
        Trip Name: Q2-MUC-TechOps
      - |

        Name : Pacific Union Services
        Category: Financial Consulting, Subscription Management
        Department: Finance
        Location: Singapore
        Amount: 129.58
        Card: Quarterly Financial Account Review
        Trip Name: unknown
      - |

        Name : FirmTrust Advisory
        Category: Legal Services, Financial Planning
        Department: Executive
        Location: London, UK
        Amount: 1534.76
        Card: Global Expansion Strategy
        Trip Name: unknown
  - source_sentence: |

      Name : ComplyTech Solutions
      Category: Regulatory Software, Consultancy Services
      Department: Compliance
      Location: Brussels, Belgium
      Amount: 1095.45
      Card: Regulatory Compliance Optimization Plan
      Trip Name: unknown
    sentences:
      - |

        Name : TechXperts Global
        Category: IT Services, Consulting
        Department: IT Operations
        Location: Berlin, Germany
        Amount: 987.49
        Card: Quarterly System Assessment
        Trip Name: unknown
      - |

        Name : Optix Global
        Category: Digital Storage Solutions, Office Essentials Provider
        Department: All Departments
        Location: Tokyo, Japan
        Amount: 568.77
        Card: Monthly Office Needs
        Trip Name: unknown
      - |

        Name : Gandalf
        Category: Financial Services, Consulting
        Department: Finance
        Location: Singapore
        Amount: 457.29
        Card: Financial Advisory Services
        Trip Name: unknown
pipeline_tag: sentence-similarity
library_name: sentence-transformers
metrics:
  - cosine_accuracy
  - dot_accuracy
  - manhattan_accuracy
  - euclidean_accuracy
  - max_accuracy
model-index:
  - name: SentenceTransformer based on BAAI/bge-base-en
    results:
      - task:
          type: triplet
          name: Triplet
        dataset:
          name: bge base en train
          type: bge-base-en-train
        metrics:
          - type: cosine_accuracy
            value: 0.8076923076923077
            name: Cosine Accuracy
          - type: dot_accuracy
            value: 0.19230769230769232
            name: Dot Accuracy
          - type: manhattan_accuracy
            value: 0.8076923076923077
            name: Manhattan Accuracy
          - type: euclidean_accuracy
            value: 0.8076923076923077
            name: Euclidean Accuracy
          - type: max_accuracy
            value: 0.8076923076923077
            name: Max Accuracy
      - task:
          type: triplet
          name: Triplet
        dataset:
          name: bge base en eval
          type: bge-base-en-eval
        metrics:
          - type: cosine_accuracy
            value: 0.9848484848484849
            name: Cosine Accuracy
          - type: dot_accuracy
            value: 0.015151515151515152
            name: Dot Accuracy
          - type: manhattan_accuracy
            value: 1
            name: Manhattan Accuracy
          - type: euclidean_accuracy
            value: 0.9848484848484849
            name: Euclidean Accuracy
          - type: max_accuracy
            value: 1
            name: Max Accuracy

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("labdmitriy/finetuned-bge-base-en")
# Run inference
sentences = [
    '\nName : ComplyTech Solutions\nCategory: Regulatory Software, Consultancy Services\nDepartment: Compliance\nLocation: Brussels, Belgium\nAmount: 1095.45\nCard: Regulatory Compliance Optimization Plan\nTrip Name: unknown\n',
    '\nName : Gandalf\nCategory: Financial Services, Consulting\nDepartment: Finance\nLocation: Singapore\nAmount: 457.29\nCard: Financial Advisory Services\nTrip Name: unknown\n',
    '\nName : TechXperts Global\nCategory: IT Services, Consulting\nDepartment: IT Operations\nLocation: Berlin, Germany\nAmount: 987.49\nCard: Quarterly System Assessment\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	0.8077
dot_accuracy	0.1923
manhattan_accuracy	0.8077
euclidean_accuracy	0.8077
max_accuracy	0.8077

Triplet

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

Metric	Value
cosine_accuracy	0.9848
dot_accuracy	0.0152
manhattan_accuracy	1.0
euclidean_accuracy	0.9848
max_accuracy	1.0

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: 39.62 tokens max: 49 tokens	0: ~3.37% 1: ~3.85% 2: ~3.85% 3: ~3.37% 4: ~6.25% 5: ~4.81% 6: ~3.85% 7: ~3.37% 8: ~4.33% 9: ~3.85% 10: ~2.40% 11: ~1.92% 12: ~3.37% 13: ~3.85% 14: ~2.88% 15: ~2.40% 16: ~5.29% 17: ~5.77% 18: ~5.29% 19: ~4.33% 20: ~1.92% 21: ~4.81% 22: ~2.40% 23: ~2.40% 24: ~2.88% 25: ~4.33% 26: ~2.88%

Samples:

sentence	label
`Name : FTC Category: Regulatory Compliance Services, Business Consulting Department: Legal Location: Toronto, Canada Amount: 3594.76 Card: Annual Compliance Assessment Trip Name: unknown`	`0`
`Name : IntelliSync Integration Category: Connectivity Services, Enterprise Solutions Department: IT Operations Location: San Francisco, CA Amount: 1387.42 Card: Global Connectivity Suite Trip Name: unknown`	`1`
`Name : Omachi Meitetsu Category: Transportation Services, Travel Services Department: Sales Location: Hakkuba Japan Amount: 120.0 Card: Quarterly Travel Expenses 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: 39.12 tokens max: 46 tokens	0: ~3.85% 1: ~1.92% 2: ~9.62% 3: ~5.77% 4: ~3.85% 5: ~3.85% 7: ~3.85% 8: ~3.85% 9: ~3.85% 10: ~3.85% 11: ~3.85% 12: ~7.69% 13: ~7.69% 14: ~1.92% 15: ~3.85% 17: ~1.92% 18: ~1.92% 19: ~3.85% 21: ~1.92% 23: ~9.62% 24: ~1.92% 25: ~1.92% 26: ~7.69%

Samples:

sentence	label
`Name : NexGen Fiscal Systems Category: Financial Software Solutions, Revenue Management Services Department: Finance Location: San Francisco, CA Amount: 2749.95 Card: Q4 Revenue Optimization Initiative Trip Name: unknown`	`15`
`Name : Midnight Brasserie Category: Culinary Experience, Event Catering Department: Marketing Location: Paris, France Amount: 456.87 Card: Quarterly Team Building Trip Name: Summer Collaboration Retreat`	`5`
`Name : Zero One Category: Media Production Department: Marketing Location: New York, NY Amount: 7500.0 Card: Sales Operating Budget Trip Name: unknown`	`13`

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
bf16: 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: True
fp16: False
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	-	0.8077
5.0	65	1.0	-

Framework Versions

Python: 3.12.8
Sentence Transformers: 3.1.1
Transformers: 4.45.2
PyTorch: 2.6.0+cu124
Accelerate: 1.3.0
Datasets: 3.2.0
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}
}