akashmaggon/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("akashmaggon/bge-base-financial-matryoshka")
# Run inference
sentences = [
    'The table presents our market risk by asset category for positions accounted for at fair value or accounted for at the lower of cost or fair value, that are not included in VaR. As of December 2023, equity was at $1,562 million and debt was at $2,446 million.',
    "What are the market risk values for Goldman Sachs' equity and debt positions not included in VaR as of December 2023?",
    "What was the conclusion of the Company's review regarding the impact of the American Rescue Plan, the Consolidated Appropriations Act, 2021, and related tax provisions on its business for the fiscal year ended June 30, 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	0.6957
cosine_accuracy@3	0.8371
cosine_accuracy@5	0.8714
cosine_accuracy@10	0.9243
cosine_precision@1	0.6957
cosine_precision@3	0.279
cosine_precision@5	0.1743
cosine_precision@10	0.0924
cosine_recall@1	0.6957
cosine_recall@3	0.8371
cosine_recall@5	0.8714
cosine_recall@10	0.9243
cosine_ndcg@10	0.8105
cosine_mrr@10	0.7742
cosine_map@100	0.7773

Information Retrieval

• Dataset: dim_512
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.7
cosine_accuracy@3	0.8286
cosine_accuracy@5	0.8671
cosine_accuracy@10	0.9186
cosine_precision@1	0.7
cosine_precision@3	0.2762
cosine_precision@5	0.1734
cosine_precision@10	0.0919
cosine_recall@1	0.7
cosine_recall@3	0.8286
cosine_recall@5	0.8671
cosine_recall@10	0.9186
cosine_ndcg@10	0.809
cosine_mrr@10	0.774
cosine_map@100	0.7776

Information Retrieval

• Dataset: dim_256
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.6929
cosine_accuracy@3	0.8186
cosine_accuracy@5	0.8586
cosine_accuracy@10	0.91
cosine_precision@1	0.6929
cosine_precision@3	0.2729
cosine_precision@5	0.1717
cosine_precision@10	0.091
cosine_recall@1	0.6929
cosine_recall@3	0.8186
cosine_recall@5	0.8586
cosine_recall@10	0.91
cosine_ndcg@10	0.8017
cosine_mrr@10	0.767
cosine_map@100	0.7712

Information Retrieval

• Dataset: dim_128
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.6871
cosine_accuracy@3	0.8071
cosine_accuracy@5	0.8586
cosine_accuracy@10	0.8986
cosine_precision@1	0.6871
cosine_precision@3	0.269
cosine_precision@5	0.1717
cosine_precision@10	0.0899
cosine_recall@1	0.6871
cosine_recall@3	0.8071
cosine_recall@5	0.8586
cosine_recall@10	0.8986
cosine_ndcg@10	0.7921
cosine_mrr@10	0.7581
cosine_map@100	0.7627

Information Retrieval

• Dataset: dim_64
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.6643
cosine_accuracy@3	0.7843
cosine_accuracy@5	0.8257
cosine_accuracy@10	0.8729
cosine_precision@1	0.6643
cosine_precision@3	0.2614
cosine_precision@5	0.1651
cosine_precision@10	0.0873
cosine_recall@1	0.6643
cosine_recall@3	0.7843
cosine_recall@5	0.8257
cosine_recall@10	0.8729
cosine_ndcg@10	0.769
cosine_mrr@10	0.7358
cosine_map@100	0.7407

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: 7 tokens mean: 44.39 tokens max: 512 tokens	min: 7 tokens mean: 20.64 tokens max: 51 tokens

• Samples:

  positive	anchor
  Johnson & Johnson reported cash and cash equivalents of $21,859 million as of the end of 2023.	What was the amount of cash and cash equivalents reported by Johnson & Johnson at the end of 2023?
  Johnson & Johnson's consolidated statements of earnings for 2023 reported total net earnings of $35,153 million.	What was the total net earnings for Johnson & Johnson in 2023?
  As of December 31, 2023, short-term investments were valued at $236,118 thousand and long-term investments at $86,676 thousand.	What is the total value of short-term and long-term investments held by the company as of December 31, 2023?

• 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
• 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: 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: 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
0.8122	10	1.5779	-	-	-	-	-
0.9746	12	-	0.7388	0.7509	0.7604	0.7081	0.7579
1.6244	20	0.6572	-	-	-	-	-
1.9492	24	-	0.7612	0.7670	0.7729	0.7269	0.7705
2.4365	30	0.4661	-	-	-	-	-
2.9239	36	-	0.7623	0.7702	0.7771	0.7386	0.7758
3.2487	40	0.3774	-	-	-	-	-
3.8985	48	-	0.7627	0.7712	0.7776	0.7407	0.7773

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 3.0.1
• Transformers: 4.41.2
• PyTorch: 2.3.1+cu121
• Accelerate: 0.32.1
• 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}
}