You can deploy this model and start earning money today!

Views

No views yet

BGE base Financial Matryoshka

Name: bge-base-financial-matryoshka
Availability: InStock

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 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.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("gallantblade/bge-base-financial-matryoshka")
# Run inference
sentences = [
    "What are the terms of Delta Air Lines' agreements with its regional carriers through Delta Connection®?",
    "Delta Connection® consists of agreements with regional airlines like Endeavor Air and SkyWest Airlines to operate flights under Delta's code. Delta controls major operational aspects like scheduling and pricing, while the regional carriers supply the services. The agreements typically last at least ten years with options for extensions.",
    'Our invention of the GPU in 1999 defined modern computer graphics and established NVIDIA as the leader in computer graphics.',
]
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

Datasets: dim_768, dim_512, dim_256, dim_128 and dim_64
Evaluated with InformationRetrievalEvaluator

Metric	dim_768	dim_512	dim_256	dim_128	dim_64
cosine_accuracy@1	0.7086	0.7057	0.7	0.6886	0.6443
cosine_accuracy@3	0.8357	0.8371	0.8314	0.8186	0.79
cosine_accuracy@5	0.8829	0.8743	0.8671	0.8557	0.8257
cosine_accuracy@10	0.9314	0.9271	0.9214	0.9229	0.8829
cosine_precision@1	0.7086	0.7057	0.7	0.6886	0.6443
cosine_precision@3	0.2786	0.279	0.2771	0.2729	0.2633
cosine_precision@5	0.1766	0.1749	0.1734	0.1711	0.1651
cosine_precision@10	0.0931	0.0927	0.0921	0.0923	0.0883
cosine_recall@1	0.7086	0.7057	0.7	0.6886	0.6443
cosine_recall@3	0.8357	0.8371	0.8314	0.8186	0.79
cosine_recall@5	0.8829	0.8743	0.8671	0.8557	0.8257
cosine_recall@10	0.9314	0.9271	0.9214	0.9229	0.8829
cosine_ndcg@10	0.8189	0.8164	0.8119	0.8032	0.7644
cosine_mrr@10	0.783	0.781	0.7769	0.7655	0.7265
cosine_map@100	0.7856	0.7839	0.7801	0.7684	0.7313

Training Details

Training Dataset

json

Dataset: json
Size: 6,300 training samples
Columns: anchor and positive
Approximate statistics based on the first 1000 samples:
anchor positive
type string string
details
min: 9 tokens
mean: 20.65 tokens
max: 45 tokens
min: 9 tokens
mean: 45.78 tokens
max: 512 tokens

Samples:

anchor	positive
`What year was Eli Lilly and Company incorporated, and in which state did this occur?`	`Eli Lilly and Company was incorporated in 1901 in Indiana to succeed the drug manufacturing business founded in Indianapolis, Indiana, in 1876 by Colonel Eli Lilly.`
`How are financial statement indexes presented in a document?`	`The financial statement indexes, including those for schedules, are organized under Part IV Item 15, specific as 'Exhibits, Financial Statement Schedules'.`
`How many physicians are part of the domestic Office of the Chief Medical Officer at DaVita as of December 31, 2023?`	`As of December 31, 2023, our domestic Chief Medical Officer leads a team of 22 nephrologists in our physician leadership team as part of our domestic Office of the Chief Medical Officer.`

Loss: MatryoshkaLoss with these parameters:
json
1{ 2 "loss": "MultipleNegativesRankingLoss", 3 "matryoshka_dims": [ 4 768, 5 512, 6 256, 7 128, 8 64 9 ], 10 "matryoshka_weights": [ 11 1, 12 1, 13 1, 14 1, 15 1 16 ], 17 "n_dims_per_step": -1 18}

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
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
0.8122	10	1.6125	-	-	-	-	-
0.9746	12	-	0.8085	0.8074	0.7977	0.7790	0.7402
1.6244	20	0.6341	-	-	-	-	-
1.9492	24	-	0.8188	0.8155	0.8081	0.7995	0.7529
2.4365	30	0.4735	-	-	-	-	-
2.9239	36	-	0.8197	0.8161	0.8107	0.8003	0.7632
3.2487	40	0.376	-	-	-	-	-
3.8985	48	-	0.8189	0.8164	0.8119	0.8032	0.7644

The bold row denotes the saved checkpoint.

Framework Versions

Python: 3.12.4
Sentence Transformers: 3.3.1
Transformers: 4.41.2
PyTorch: 2.4.0+cu124
Accelerate: 1.1.1
Datasets: 2.19.1
Tokenizers: 0.19.1

Citation

BibTeX

Sentence Transformers

bibtex

@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

bibtex

@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

bibtex

@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}
}

	anchor	positive
type	string	string
details	min: 9 tokens mean: 20.65 tokens max: 45 tokens	min: 9 tokens mean: 45.78 tokens max: 512 tokens