RK-1235/bge-base-FIR-matryoshka-BASELINE-10epochs-finetune

BGE base Financial Matryoshka

This is a sentence-transformers model finetuned from BAAI/bge-base-en 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
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • json
• 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("RK-1235/bge-base-FIR-matryoshka-BASELINE-10epochs-finetune")
# Run inference
sentences = [
    'Item 8. Financial Statements and Supplementary Data. The Consolidated Financial Statements, together with the Notes thereto and the report thereon dated February 16, 2024, of PricewaterhouseCoopers LLP, the Firm’s independent registered public accounting firm (PCAOB ID 238).',
    'What type of data does Item 8 in a financial document contain?',
    "How did the assumptions and estimates used for assessing the fair value of reporting units potentially impact the company's financial statements?",
]
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 with these parameters:

  {
      "truncate_dim": 768
  }
  

Metric	Value
cosine_accuracy@1	0.1672
cosine_accuracy@3	0.3292
cosine_accuracy@5	0.3927
cosine_accuracy@10	0.4788
cosine_precision@1	0.1672
cosine_precision@3	0.1097
cosine_precision@5	0.0785
cosine_precision@10	0.0479
cosine_recall@1	0.1672
cosine_recall@3	0.3292
cosine_recall@5	0.3927
cosine_recall@10	0.4788
cosine_ndcg@10	0.3177
cosine_mrr@10	0.2669
cosine_map@100	0.2763

Information Retrieval

• Dataset: dim_512
• Evaluated with InformationRetrievalEvaluator with these parameters:

  {
      "truncate_dim": 512
  }
  

Metric	Value
cosine_accuracy@1	0.1693
cosine_accuracy@3	0.3241
cosine_accuracy@5	0.3803
cosine_accuracy@10	0.4701
cosine_precision@1	0.1693
cosine_precision@3	0.108
cosine_precision@5	0.0761
cosine_precision@10	0.047
cosine_recall@1	0.1693
cosine_recall@3	0.3241
cosine_recall@5	0.3803
cosine_recall@10	0.4701
cosine_ndcg@10	0.3134
cosine_mrr@10	0.2642
cosine_map@100	0.2739

Information Retrieval

• Dataset: dim_256
• Evaluated with InformationRetrievalEvaluator with these parameters:

  {
      "truncate_dim": 256
  }
  

Metric	Value
cosine_accuracy@1	0.1547
cosine_accuracy@3	0.2934
cosine_accuracy@5	0.3547
cosine_accuracy@10	0.4292
cosine_precision@1	0.1547
cosine_precision@3	0.0978
cosine_precision@5	0.0709
cosine_precision@10	0.0429
cosine_recall@1	0.1547
cosine_recall@3	0.2934
cosine_recall@5	0.3547
cosine_recall@10	0.4292
cosine_ndcg@10	0.2866
cosine_mrr@10	0.2417
cosine_map@100	0.2524

Information Retrieval

• Dataset: dim_128
• Evaluated with InformationRetrievalEvaluator with these parameters:

  {
      "truncate_dim": 128
  }
  

Metric	Value
cosine_accuracy@1	0.1248
cosine_accuracy@3	0.2533
cosine_accuracy@5	0.3022
cosine_accuracy@10	0.3715
cosine_precision@1	0.1248
cosine_precision@3	0.0844
cosine_precision@5	0.0604
cosine_precision@10	0.0372
cosine_recall@1	0.1248
cosine_recall@3	0.2533
cosine_recall@5	0.3022
cosine_recall@10	0.3715
cosine_ndcg@10	0.243
cosine_mrr@10	0.2026
cosine_map@100	0.213

Information Retrieval

• Dataset: dim_64
• Evaluated with InformationRetrievalEvaluator with these parameters:

  {
      "truncate_dim": 64
  }
  

Metric	Value
cosine_accuracy@1	0.092
cosine_accuracy@3	0.181
cosine_accuracy@5	0.2234
cosine_accuracy@10	0.2949
cosine_precision@1	0.092
cosine_precision@3	0.0603
cosine_precision@5	0.0447
cosine_precision@10	0.0295
cosine_recall@1	0.092
cosine_recall@3	0.181
cosine_recall@5	0.2234
cosine_recall@10	0.2949
cosine_ndcg@10	0.1842
cosine_mrr@10	0.15
cosine_map@100	0.1594

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: 46.06 tokens max: 371 tokens	min: 8 tokens mean: 20.8 tokens max: 51 tokens

• Samples:

  positive	anchor
  As of December 31, 2023, a 5 percent change in the contingent consideration liabilities would result in a change in income before income taxes of $5.2 million.	How would a 5% change in the contingent consideration liabilities impact income before taxes as of December 31, 2023?
  NIKE, Inc.'s principal business activity involves the design, development, and worldwide marketing and selling of athletic footwear, apparel, equipment, accessories, and services.	What is the principal business activity of NIKE, Inc.?
  During 2023, changes in foreign currencies relative to the U.S. dollar negatively impacted net sales by approximately $3,484, 156 basis points, compared to 2022, attributable to our Canadian and Other International operations.	What was the overall impact of foreign currencies on net sales in 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: 8
• learning_rate: 2e-05
• num_train_epochs: 10
• 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: 8
• 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: 10
• 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: None
• hub_always_push: False
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_inputs_for_metrics: False
• include_for_metrics: []
• 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
• 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
• average_tokens_across_devices: False
• prompts: None
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	dim_768_cosine_ndcg@10	dim_512_cosine_ndcg@10	dim_256_cosine_ndcg@10	dim_128_cosine_ndcg@10	dim_64_cosine_ndcg@10
0.4061	10	47.317	-	-	-	-	-
0.8122	20	29.4505	-	-	-	-	-
1.0	25	-	0.3433	0.334	0.3129	0.2614	0.1806
1.2030	30	14.0234	-	-	-	-	-
1.6091	40	8.2499	-	-	-	-	-
2.0	50	5.4979	0.3146	0.3087	0.2851	0.2389	0.1790
2.4061	60	3.9809	-	-	-	-	-
2.8122	70	3.5321	-	-	-	-	-
3.0	75	-	0.3246	0.3183	0.2928	0.2412	0.1836
3.2030	80	2.7593	-	-	-	-	-
3.6091	90	2.4589	-	-	-	-	-
4.0	100	2.5858	0.3270	0.3195	0.2987	0.2452	0.1843
4.4061	110	2.1241	-	-	-	-	-
4.8122	120	1.7721	-	-	-	-	-
5.0	125	-	0.3167	0.3128	0.2880	0.2430	0.1862
5.2030	130	2.0458	-	-	-	-	-
5.6091	140	1.8376	-	-	-	-	-
6.0	150	1.7751	0.3123	0.3065	0.2851	0.2412	0.1825
6.4061	160	1.6278	-	-	-	-	-
6.8122	170	1.8976	-	-	-	-	-
7.0	175	-	0.3154	0.3092	0.2875	0.2428	0.1846
7.2030	180	1.582	-	-	-	-	-
7.6091	190	1.4319	-	-	-	-	-
8.0	200	1.4672	0.3170	0.3123	0.2862	0.2437	0.1841
8.4061	210	1.7736	-	-	-	-	-
8.8122	220	1.4284	-	-	-	-	-
9.0	225	-	0.3194	0.3120	0.2877	0.2423	0.1832
9.2030	230	1.1812	-	-	-	-	-
9.6091	240	1.4361	-	-	-	-	-
10.0	250	1.5928	0.3177	0.3134	0.2866	0.2430	0.1842

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 4.1.0
• Transformers: 4.52.2
• PyTorch: 2.6.0+cu124
• Accelerate: 1.7.0
• Datasets: 3.6.0
• Tokenizers: 0.21.1

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}
}