ybWw/bge-base-financial-matryoshka

SentenceTransformer based on BAAI/bge-base-en-v1.5

This is a sentence-transformers model finetuned from BAAI/bge-base-en-v1.5 on the json 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: BAAI/bge-base-en-v1.5
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 tokens
• Similarity Function: Cosine Similarity
• Training Dataset:
  • json

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("ybWw/bge-base-financial-matryoshka")
# Run inference
sentences = [
    'A 1% increase in medical cost PMPM trend factors led to an increase in medical costs payable by $1,128 million for the most recent two months as of December 31, 2023.',
    'What financial impact would a 1% increase in medical cost PMPM trend factors have on medical costs payable for the most recent two months as of December 31, 2023?',
    'How does Kroger present its financial performance in its reporting?',
]
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

Information Retrieval

• Dataset: dim_768
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.6843
cosine_accuracy@3	0.8186
cosine_accuracy@5	0.8571
cosine_accuracy@10	0.9
cosine_precision@1	0.6843
cosine_precision@3	0.2729
cosine_precision@5	0.1714
cosine_precision@10	0.09
cosine_recall@1	0.6843
cosine_recall@3	0.8186
cosine_recall@5	0.8571
cosine_recall@10	0.9
cosine_ndcg@10	0.7945
cosine_mrr@10	0.7606
cosine_map@100	0.7643

Information Retrieval

• Dataset: dim_512
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.6771
cosine_accuracy@3	0.8129
cosine_accuracy@5	0.8571
cosine_accuracy@10	0.9
cosine_precision@1	0.6771
cosine_precision@3	0.271
cosine_precision@5	0.1714
cosine_precision@10	0.09
cosine_recall@1	0.6771
cosine_recall@3	0.8129
cosine_recall@5	0.8571
cosine_recall@10	0.9
cosine_ndcg@10	0.7898
cosine_mrr@10	0.7544
cosine_map@100	0.7579

Information Retrieval

• Dataset: dim_256
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.6829
cosine_accuracy@3	0.8129
cosine_accuracy@5	0.8557
cosine_accuracy@10	0.89
cosine_precision@1	0.6829
cosine_precision@3	0.271
cosine_precision@5	0.1711
cosine_precision@10	0.089
cosine_recall@1	0.6829
cosine_recall@3	0.8129
cosine_recall@5	0.8557
cosine_recall@10	0.89
cosine_ndcg@10	0.789
cosine_mrr@10	0.7562
cosine_map@100	0.7601

Information Retrieval

• Dataset: dim_128
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.6657
cosine_accuracy@3	0.8
cosine_accuracy@5	0.8314
cosine_accuracy@10	0.88
cosine_precision@1	0.6657
cosine_precision@3	0.2667
cosine_precision@5	0.1663
cosine_precision@10	0.088
cosine_recall@1	0.6657
cosine_recall@3	0.8
cosine_recall@5	0.8314
cosine_recall@10	0.88
cosine_ndcg@10	0.7745
cosine_mrr@10	0.7406
cosine_map@100	0.7444

Information Retrieval

• Dataset: dim_64
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.6357
cosine_accuracy@3	0.7686
cosine_accuracy@5	0.8086
cosine_accuracy@10	0.8514
cosine_precision@1	0.6357
cosine_precision@3	0.2562
cosine_precision@5	0.1617
cosine_precision@10	0.0851
cosine_recall@1	0.6357
cosine_recall@3	0.7686
cosine_recall@5	0.8086
cosine_recall@10	0.8514
cosine_ndcg@10	0.7447
cosine_mrr@10	0.7103
cosine_map@100	0.7149

Training Details

Training Dataset

json

• Dataset: json
• Size: 6,300 training samples
• Columns: positive and anchor
• Approximate statistics based on the first 1000 samples:

  	positive	anchor
  type	string	string
  details	min: 6 tokens mean: 46.17 tokens max: 272 tokens	min: 8 tokens mean: 20.72 tokens max: 42 tokens

• Samples:

  positive	anchor
  For fiscal 2023, the net cash provided by operating activities was $7,111 million.	What was the net cash provided by operating activities in fiscal 2023?
  Penalties for impermissible use or disclosure of PHI were increased by the HITECH Act by imposing tiered penalties of more than $50,000 per violation and up to $1.5 million per year for identical violations.	What are the consequences of impermissible use or disclosure of PHI according to the HITECH Act?
  Global Day of Joy is Hasbro’s annual, company-wide day of service and has become a cherished tradition. Global Day of Joy takes place every December, and employees from each Hasbro office participate in service projects to benefit a variety of organizations.	What is the purpose of the Global Day of Joy at Hasbro?

• Loss: MatryoshkaLoss with these parameters:

  {
      "loss": "MultipleNegativesRankingLoss",
      "matryoshka_dims": [
          768,
          512,
          256,
          128,
          64
      ],
      "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: 4
• 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
• 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: 4
• 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: 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
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	dim_768_cosine_map@100	dim_512_cosine_map@100	dim_256_cosine_map@100	dim_128_cosine_map@100	dim_64_cosine_map@100
0.8122	10	1.5383	-	-	-	-	-
0.9746	12	-	0.7575	0.7597	0.7428	0.7303	0.6889
1.6244	20	0.6278	-	-	-	-	-
1.9492	24	-	0.7623	0.7580	0.7582	0.7417	0.7100
2.4365	30	0.4388	-	-	-	-	-
2.9239	36	-	0.7649	0.7576	0.7571	0.7465	0.7142
3.2487	40	0.3729	-	-	-	-	-
3.8985	48	-	0.7643	0.7579	0.7601	0.7444	0.7149

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.8.20
• Sentence Transformers: 3.2.1
• Transformers: 4.41.2
• PyTorch: 2.1.2+cu121
• Accelerate: 1.0.1
• Datasets: 2.19.1
• Tokenizers: 0.19.1

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}
}