kperkins411/bge-base-financial-matryoshka

BGE base Financial Matryoshka

This is a sentence-transformers model finetuned from BAAI/bge-base-en-v1.5. 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
• Language: en
• 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': 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("kperkins411/bge-base-financial-matryoshka")
# Run inference
sentences = [
    'Alternative Payments Providers: These providers, such as closed commerce ecosystems, BNPL solutions and cryptocurrency platforms, often have a primary focus of enabling payments through ecommerce and mobile channels; however, they are expanding or may expand their offerings to the physical point of sale. These companies may process payments using in-house account transfers between parties, electronic funds transfer networks like the ACH, global or local networks like Visa, or some combination of the foregoing.',
    'What are some examples of alternative payments providers and how do they compete with Visa?',
    "How much did the company's currently payable U.S. taxes amount to in 2023?",
]
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	0.6886
cosine_accuracy@3	0.8329
cosine_accuracy@5	0.8743
cosine_accuracy@10	0.9143
cosine_precision@1	0.6886
cosine_precision@3	0.2776
cosine_precision@5	0.1749
cosine_precision@10	0.0914
cosine_recall@1	0.6886
cosine_recall@3	0.8329
cosine_recall@5	0.8743
cosine_recall@10	0.9143
cosine_ndcg@10	0.8045
cosine_mrr@10	0.769
cosine_map@100	0.7722

Information Retrieval

• Dataset: dim_512
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.6971
cosine_accuracy@3	0.8343
cosine_accuracy@5	0.8743
cosine_accuracy@10	0.9071
cosine_precision@1	0.6971
cosine_precision@3	0.2781
cosine_precision@5	0.1749
cosine_precision@10	0.0907
cosine_recall@1	0.6971
cosine_recall@3	0.8343
cosine_recall@5	0.8743
cosine_recall@10	0.9071
cosine_ndcg@10	0.8044
cosine_mrr@10	0.7713
cosine_map@100	0.775

Information Retrieval

• Dataset: dim_256
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.6914
cosine_accuracy@3	0.8257
cosine_accuracy@5	0.8714
cosine_accuracy@10	0.91
cosine_precision@1	0.6914
cosine_precision@3	0.2752
cosine_precision@5	0.1743
cosine_precision@10	0.091
cosine_recall@1	0.6914
cosine_recall@3	0.8257
cosine_recall@5	0.8714
cosine_recall@10	0.91
cosine_ndcg@10	0.8034
cosine_mrr@10	0.7691
cosine_map@100	0.7725

Information Retrieval

• Dataset: dim_128
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.6743
cosine_accuracy@3	0.81
cosine_accuracy@5	0.8543
cosine_accuracy@10	0.9
cosine_precision@1	0.6743
cosine_precision@3	0.27
cosine_precision@5	0.1709
cosine_precision@10	0.09
cosine_recall@1	0.6743
cosine_recall@3	0.81
cosine_recall@5	0.8543
cosine_recall@10	0.9
cosine_ndcg@10	0.7881
cosine_mrr@10	0.7522
cosine_map@100	0.756

Information Retrieval

• Dataset: dim_64
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.6386
cosine_accuracy@3	0.7671
cosine_accuracy@5	0.8243
cosine_accuracy@10	0.87
cosine_precision@1	0.6386
cosine_precision@3	0.2557
cosine_precision@5	0.1649
cosine_precision@10	0.087
cosine_recall@1	0.6386
cosine_recall@3	0.7671
cosine_recall@5	0.8243
cosine_recall@10	0.87
cosine_ndcg@10	0.7529
cosine_mrr@10	0.7155
cosine_map@100	0.7206

Training Details

Training Dataset

Unnamed Dataset

• 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: 45.51 tokens max: 371 tokens	min: 10 tokens mean: 20.83 tokens max: 45 tokens

• Samples:

  positive	anchor
  Activities related to sales before 2023 experienced adjustments due to changes in estimates, impacting the rebates and chargebacks accounts, and led to an ending balance of $4,493 million for the year 2023.	What adjustments were made to the rebates and chargebacks balances for previous years' sales and how did they affect the end of year balance in 2023?
  We’re focused on making hosting just as popular as traveling on Airbnb. We will continue to invest in growing the size and quality of our Host community. We plan to attract more Hosts globally by expanding use cases and supporting all different types of Hosts, including those who host occasionally.	What is Airbnb's long-term corporate strategy regarding hosting?
  Due to protectionist measures in various regions, Nike has experienced increased product costs. The company responds by monitoring trends, engaging in processes to mitigate restrictions, and advocating for trade liberalization in trade agreements.	What challenges related to trade protectionism has Nike faced, and what measures has the company taken in response?

• 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_128_cosine_map@100	dim_256_cosine_map@100	dim_512_cosine_map@100	dim_64_cosine_map@100	dim_768_cosine_map@100
0.96	3	-	0.7116	0.7341	0.7448	0.6550	0.7455
1.92	6	-	0.7317	0.7520	0.7586	0.6975	0.7591
2.88	9	-	0.7334	0.7553	0.7631	0.7039	0.7630
3.2	10	3.3636	-	-	-	-	-
3.84	12	-	0.7368	0.759	0.7634	0.7054	0.7638
0.96	3	-	0.7415	0.7601	0.7672	0.7102	0.7661
1.92	6	-	0.7486	0.7683	0.7720	0.7205	0.7718
2.88	9	-	0.7556	0.7718	0.7750	0.7215	0.7717
3.2	10	1.66	-	-	-	-	-
3.84	12	-	0.756	0.7725	0.775	0.7206	0.7722

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.11.9
• Sentence Transformers: 3.0.1
• Transformers: 4.41.2
• PyTorch: 2.1.2+cu121
• Accelerate: 0.31.0
• 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}
}