anishareddyalla/bge-base-financial-matryoshka-anisha

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("anishareddyalla/bge-base-financial-matryoshka-anisha")
# Run inference
sentences = [
    "In 1983, Walmart opened its first Sam's Club, and in 1988, it opened its first supercenter.",
    "When did Walmart open its first Sam's Club and supercenter?",
    'Which standards and guidelines does the company use for informing its sustainability disclosures?',
]
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	0.7029
cosine_accuracy@3	0.8371
cosine_accuracy@5	0.8729
cosine_accuracy@10	0.9186
cosine_precision@1	0.7029
cosine_precision@3	0.279
cosine_precision@5	0.1746
cosine_precision@10	0.0919
cosine_recall@1	0.7029
cosine_recall@3	0.8371
cosine_recall@5	0.8729
cosine_recall@10	0.9186
cosine_ndcg@10	0.812
cosine_mrr@10	0.7777
cosine_map@100	0.781

Information Retrieval

• Dataset: dim_512
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.6986
cosine_accuracy@3	0.8329
cosine_accuracy@5	0.8643
cosine_accuracy@10	0.9243
cosine_precision@1	0.6986
cosine_precision@3	0.2776
cosine_precision@5	0.1729
cosine_precision@10	0.0924
cosine_recall@1	0.6986
cosine_recall@3	0.8329
cosine_recall@5	0.8643
cosine_recall@10	0.9243
cosine_ndcg@10	0.8105
cosine_mrr@10	0.7743
cosine_map@100	0.7771

Information Retrieval

• Dataset: dim_256
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.6943
cosine_accuracy@3	0.8271
cosine_accuracy@5	0.8586
cosine_accuracy@10	0.9086
cosine_precision@1	0.6943
cosine_precision@3	0.2757
cosine_precision@5	0.1717
cosine_precision@10	0.0909
cosine_recall@1	0.6943
cosine_recall@3	0.8271
cosine_recall@5	0.8586
cosine_recall@10	0.9086
cosine_ndcg@10	0.8026
cosine_mrr@10	0.7687
cosine_map@100	0.7726

Information Retrieval

• Dataset: dim_128
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.6886
cosine_accuracy@3	0.8157
cosine_accuracy@5	0.8571
cosine_accuracy@10	0.9071
cosine_precision@1	0.6886
cosine_precision@3	0.2719
cosine_precision@5	0.1714
cosine_precision@10	0.0907
cosine_recall@1	0.6886
cosine_recall@3	0.8157
cosine_recall@5	0.8571
cosine_recall@10	0.9071
cosine_ndcg@10	0.7973
cosine_mrr@10	0.7622
cosine_map@100	0.7657

Information Retrieval

• Dataset: dim_64
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.66
cosine_accuracy@3	0.7986
cosine_accuracy@5	0.8357
cosine_accuracy@10	0.8829
cosine_precision@1	0.66
cosine_precision@3	0.2662
cosine_precision@5	0.1671
cosine_precision@10	0.0883
cosine_recall@1	0.66
cosine_recall@3	0.7986
cosine_recall@5	0.8357
cosine_recall@10	0.8829
cosine_ndcg@10	0.7716
cosine_mrr@10	0.7361
cosine_map@100	0.7401

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: 8 tokens mean: 46.43 tokens max: 439 tokens	min: 8 tokens mean: 20.76 tokens max: 43 tokens

• Samples:

  positive	anchor
  The Company’s human capital management strategy is built on three fundamental focus areas: Attracting and recruiting the best talent, Developing and retaining talent, Empowering and inspiring talent.	What strategies are outlined in the Company's human capital management?
  Opinion on the Consolidated Financial Statements We have audited the accompanying consolidated balance sheets of Costco Wholesale Corporation and subsidiaries (the Company) as of September 3, 2023, and August 28, 2022, the related consolidated statements of income, comprehensive income, equity, and cash flows for the 53-week period ended September 3, 2023, and the 52-week periods ended August 28, 2022, and August 29, 2021, and the related notes (collectively, the consolidated financial statements). In our opinion, the consolidated financial statements present fairly, in all material respects, the financial position of the Company as of September 3, 2023, and August 28, 2022, and the results of its operations and its cash flows for each of the 53-week period ended September 3, 2023, and the 52-week periods ended August 28, 2022, and August 29, 2021, in conformity with U.S. generally accepted accounting principles.	What was the opinion of the independent registered public accounting firm on Costco Wholesale Corporation's consolidated financial statements for the year ended September 3, 2023?
  Nonperforming loans and leases are generally those that have been placed on nonaccrual status, such as when they are 90 days past due or have confirmed cases of fraud or bankruptcy. Additionally, specific types of loans like consumer real estate-secured loans are classified as nonperforming at 90 days past due unless they are fully insured, and commercial loans and leases are classified as nonperforming when past due 90 days or more unless well-secured and in the process of collection.	What criteria are used to classify loans and leases as nonperforming according to the described credit policy?

• 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
• eval_on_start: 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	1.5488	-	-	-	-	-
0.9746	12	-	0.7540	0.7565	0.7660	0.7176	0.7693
1.6244	20	0.674	-	-	-	-	-
1.9492	24	-	0.7622	0.7715	0.7781	0.7352	0.7790
2.4365	30	0.4592	-	-	-	-	-
2.9239	36	-	0.7648	0.7729	0.7778	0.7384	0.7799
3.2487	40	0.4113	-	-	-	-	-
3.8985	48	-	0.7657	0.7726	0.7771	0.7401	0.7810

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 3.0.1
• Transformers: 4.42.4
• PyTorch: 2.3.1+cu121
• Accelerate: 0.32.1
• Datasets: 2.20.0
• 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}
}