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 : Gandalf
      Category: Financial Services, Consulting
      Department: Finance
      Location: Singapore
      Amount: 457.29
      Card: Financial Advisory Services
      Trip Name: unknown
    sentences:
      - |

        Name : InterGlobal Tech
        Category: Business Software Solutions, Data Processing Services
        Department: Marketing
        Location: New York, NY
        Amount: 1249.95
        Card: Marketing Automation Tools
        Trip Name: unknown
      - |

        Name : Nuvotek Solutions
        Category: Consulting Services, Managed IT Services
        Department: Information Security
        Location: Berlin, Germany
        Amount: 879.65
        Card: Annual Cybersecurity Resilience Program
        Trip Name: unknown
      - |

        Name : Omega Systems Inc.
        Category: Integrated Business Solutions, Enterprise Software Sales
        Department: Research & Development
        Location: Oslo, Norway
        Amount: 1943.75
        Card: AI Development Suite
        Trip Name: unknown
  - source_sentence: |

      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
    sentences:
      - |

        Name : GlobalRes Workforce Solutions
        Category: Remote Work Platforms, HR Technology Vendors
        Department: Engineering
        Location: Barcelona, Spain
        Amount: 1894.27
        Card: Hybrid Work Enablement
        Trip Name: unknown
      - |

        Name : InterLang Solutions
        Category: Language Interpretation Services, Remote Collaboration Tools
        Department: HR
        Location: Tokyo, Japan
        Amount: 1642.59
        Card: Diversity & Inclusion Initiatives
        Trip Name: unknown
      - |

        Name : CovaRisk Consulting
        Category: Risk Advisory, Financial Services
        Department: Legal
        Location: Toronto, Canada
        Amount: 1124.37
        Card: Assurance Payment
        Trip Name: unknown
  - source_sentence: |

      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
    sentences:
      - |

        Name : Digital Wave Solutions
        Category: IT Infrastructure Services, Data Analytic Platforms
        Department: Finance
        Location: San Francisco, CA
        Amount: 1748.92
        Card: Annual Data Management & Reporting
        Trip Name: unknown
      - |

        Name : Analytix Global Solutions
        Category: Business Intelligence Services, Regulatory Compliance Tools
        Department: Finance
        Location: London, UK
        Amount: 1323.67
        Card: Financial Compliance Enhancement
        Trip Name: unknown
      - |

        Name : Daesung Enterprises
        Category: Catering Services, Event Management
        Department: Sales
        Location: Lisbon, Portugal
        Amount: 375.45
        Card: Q4 Client Engagement Events
        Trip Name: unknown
  - source_sentence: |

      Name : Kanzan Solutions
      Category: Consulting Services, Business Advisory
      Department: Legal
      Location: Tokyo, Japan
      Amount: 3900.75
      Card: Quarterly Compliance Review
      Trip Name: unknown
    sentences:
      - |

        Name : Alta Via Mix
        Category: Airline Catering, Luxury Travel Services
        Department: Executive
        Location: Milan, Italy
        Amount: 1925.49
        Card: Executive Incentive Program
        Trip Name: Annual Leadership Summit
      - |

        Name : RBS
        Category: Financial Services, Business Consultancy
        Department: Finance
        Location: Toronto, Canada
        Amount: 1134.28
        Card: Cross-Border Transaction Facilitation
        Trip Name: unknown
      - |

        Name : InnovaThink Global
        Category: Management Consultancy, Technical Training Services
        Department: HR
        Location: Zurich, Switzerland
        Amount: 1675.32
        Card: Innovation and Efficiency Program
        Trip Name: unknown
  - source_sentence: |

      Name : NetWise Solutions
      Category: Data Transfer Services, Digital Infrastructure
      Department: Product
      Location: Singapore
      Amount: 1579.42
      Card: Global Network Enhancement
      Trip Name: unknown
    sentences:
      - |

        Name : Fernández & Co. Services
        Category: Property Management, Facility Services
        Department: Office Administration
        Location: Madrid, Spain
        Amount: 1245.67
        Card: Monthly Facility Operations
        Trip Name: unknown
      - |

        Name : AeroDyn Research
        Category: Research Services, Data Analysis
        Department: Research & Development
        Location: Amsterdam, Netherlands
        Amount: 2457.42
        Card: Annual Innovation Assessment
        Trip Name: unknown
      - |

        Name : Global Horizon Travel
        Category: Travel Services, Package Deals
        Department: Sales
        Location: Tokyo, Japan
        Amount: 1199.75
        Card: Annual Sales Retreat
        Trip Name: Sales Strategy Summit
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: 0.8605769230769231
            name: Cosine Accuracy
          - type: dot_accuracy
            value: 0.13942307692307693
            name: Dot Accuracy
          - type: manhattan_accuracy
            value: 0.8413461538461539
            name: Manhattan Accuracy
          - type: euclidean_accuracy
            value: 0.8605769230769231
            name: Euclidean Accuracy
          - type: max_accuracy
            value: 0.8605769230769231
            name: Max Accuracy
      - task:
          type: triplet
          name: Triplet
        dataset:
          name: bge base en eval
          type: bge-base-en-eval
        metrics:
          - type: cosine_accuracy
            value: 0.9242424242424242
            name: Cosine Accuracy
          - type: dot_accuracy
            value: 0.07575757575757576
            name: Dot Accuracy
          - type: manhattan_accuracy
            value: 0.9545454545454546
            name: Manhattan Accuracy
          - type: euclidean_accuracy
            value: 0.9242424242424242
            name: Euclidean Accuracy
          - type: max_accuracy
            value: 0.9545454545454546
            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("dshvadskiy/finetuned-bge-base-en")
# Run inference
sentences = [
    '\nName : NetWise Solutions\nCategory: Data Transfer Services, Digital Infrastructure\nDepartment: Product\nLocation: Singapore\nAmount: 1579.42\nCard: Global Network Enhancement\nTrip Name: unknown\n',
    '\nName : Global Horizon Travel\nCategory: Travel Services, Package Deals\nDepartment: Sales\nLocation: Tokyo, Japan\nAmount: 1199.75\nCard: Annual Sales Retreat\nTrip Name: Sales Strategy Summit\n',
    '\nName : AeroDyn Research\nCategory: Research Services, Data Analysis\nDepartment: Research & Development\nLocation: Amsterdam, Netherlands\nAmount: 2457.42\nCard: Annual Innovation 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	0.8606
dot_accuracy	0.1394
manhattan_accuracy	0.8413
euclidean_accuracy	0.8606
max_accuracy	0.8606

Triplet

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

Metric	Value
cosine_accuracy	0.9242
dot_accuracy	0.0758
manhattan_accuracy	0.9545
euclidean_accuracy	0.9242
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: 32 tokens mean: 39.5 tokens max: 49 tokens	0: ~5.29% 1: ~4.81% 2: ~3.37% 3: ~3.85% 4: ~3.85% 5: ~5.77% 6: ~1.92% 7: ~2.88% 8: ~5.29% 9: ~5.29% 10: ~4.33% 11: ~2.40% 12: ~3.85% 13: ~2.88% 14: ~4.33% 15: ~3.37% 16: ~3.37% 17: ~1.44% 18: ~4.33% 19: ~4.81% 20: ~3.85% 21: ~2.88% 22: ~5.77% 23: ~3.37% 24: ~2.88% 25: ~0.96% 26: ~2.88%

Samples:

sentence	label
`Name : Yijie Logistics Category: Logistics Services Department: Sales Location: Berlin, Germany Amount: 485.67 Card: Quarterly Client Visit and Logistics Coordination Trip Name: unknown`	`0`
`Name : Serenity Solutions Category: Office Wellness Solutions Department: Office Administration Location: Munich, Germany Amount: 772.58 Card: Ergonomic Office Enhancements Trip Name: unknown`	`1`
`Name : Cortec International Category: Event Management Services, Business Solutions Department: Sales Location: London, UK Amount: 1337.25 Card: Global Sales Summit Participation 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: 34 tokens mean: 39.62 tokens max: 46 tokens	0: ~3.85% 3: ~1.92% 4: ~5.77% 5: ~5.77% 6: ~3.85% 7: ~1.92% 8: ~1.92% 9: ~1.92% 10: ~3.85% 11: ~9.62% 12: ~5.77% 13: ~3.85% 14: ~1.92% 15: ~9.62% 17: ~1.92% 18: ~3.85% 20: ~1.92% 21: ~9.62% 22: ~1.92% 23: ~3.85% 24: ~1.92% 25: ~5.77% 26: ~7.69%

Samples:

sentence	label
`Name : Versatile Systems Ltd. Category: Office Management Solutions, Software Solutions Department: Office Administration Location: Tokyo, Japan Amount: 845.67 Card: Integrated Office Infrastructure Trip Name: unknown`	`21`
`Name : NexGen Comms Category: Telecom Services, Communications Solutions Department: Sales Location: Berlin, Germany Amount: 879.45 Card: Q2 Client Outreach Program Trip Name: unknown`	`23`
`Name : Digital Wave Solutions Category: IT Infrastructure Services, Data Analytic Platforms Department: Finance Location: San Francisco, CA Amount: 1748.92 Card: Annual Data Management & Reporting Trip Name: unknown`	`18`

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
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: 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	-	0.8606
5.0	65	0.9545	-

Framework Versions

Python: 3.9.16
Sentence Transformers: 3.1.1
Transformers: 4.45.2
PyTorch: 2.6.0
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}
}