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-10ep")
# Run inference
sentences = [
    "on two motorcycles. The informant pointed out that the person wearing jeans and vest was the one. The police waited standing in the shade of Karra Power House for the approaching person. The informant left after pointing. When the person came within the police's sight, they caught him about 20 steps away from Karra Power House with a single warning shot. Upon searching the arrested person and asking for his name and address, he said his name was Sikandar Khan, age around 32 years, residing in Mo Baradari, Shah0 Sadar Bazaar Police Station. During the search conducted by No. 2156, a country-made 315 bore gun was found sewn inside the right pocket of the jeans he was wearing, with a barrel length of 11 inches and a stock length of 7 inches, having a hammer trigger attached, the trigger guard being aluminum, and a plate attached to open and close the barrel with a spring of 4 inches attached on both sides with girlie caps screwed with two screws. From the left pocket of the jeans, one",
    'What was the name and age of the person caught by the police near Karra Power House?',
    'Where did Monica SLO Jayprakash say she was calling from when she spoke to her mother?',
]
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]

Training Details

Training Dataset

json

Dataset: json
Size: 5,251 training samples
Columns: positive and anchor
Approximate statistics based on the first 1000 samples:
positive anchor
type string string
details
min: 14 tokens
mean: 211.86 tokens
max: 391 tokens

min: 11 tokens
mean: 19.91 tokens
max: 52 tokens

	positive	anchor
type	string	string
details	min: 14 tokens mean: 211.86 tokens max: 391 tokens	min: 11 tokens mean: 19.91 tokens max: 52 tokens

Samples:

positive	anchor
First Information Report: Officer Thanadhiksh Thahbarpur Ajmer respectfully submits that the petitioner Manju Yadav states that my husband Santosh Yadav, son Sw. Harikesh Yadav, brother-in-law Mayank Yadav, son Sw. Heetikesh Yadav, and mother-in-law Radhika, wife of Sw. Harikesh Yadav, these three people have beaten me together. If the villagers had not intervened, these three would have killed me and my son. The constant intention of my mother-in-law Radhika, wife of Sw. Harikesh Yadav, is to beat me to death. My husband Santosh Yadav and brother-in-law Mayank Yadav have been indecent with me. My husband cut off my ear while my brother-in-law Mayank Yadav tore my sari and tried to undress me, which is why I was trying to protect myself with bricks. If the villagers had not intervened, these people would have undressed and killed me. These three come every day asking for dowry from me and beat me, insult me, and abuse me. There is a danger to my life and my son's life from these three	`Who are the three individuals mentioned in the report who have beaten Manju Yadav?`
First Information Report: Officer Thanadhiksh Thahbarpur Ajmer respectfully submits that the petitioner Manju Yadav states that my husband Santosh Yadav, son Sw. Harikesh Yadav, brother-in-law Mayank Yadav, son Sw. Heetikesh Yadav, and mother-in-law Radhika, wife of Sw. Harikesh Yadav, these three people have beaten me together. If the villagers had not intervened, these three would have killed me and my son. The constant intention of my mother-in-law Radhika, wife of Sw. Harikesh Yadav, is to beat me to death. My husband Santosh Yadav and brother-in-law Mayank Yadav have been indecent with me. My husband cut off my ear while my brother-in-law Mayank Yadav tore my sari and tried to undress me, which is why I was trying to protect myself with bricks. If the villagers had not intervened, these people would have undressed and killed me. These three come every day asking for dowry from me and beat me, insult me, and abuse me. There is a danger to my life and my son's life from these three	`What actions did Manju Yadav's husband Santosh Yadav and brother-in-law Mayank Yadav take against her?`
`to my life and my son's life from these three people. Therefore, I request you to grant justice. Petitioner Manju Yadav, husband Santosh Yadav, village Sikandarpur, post Dharmdaspur, Nawada district-Ajmer, mobile number-8423229252, date-16/02/2024 HMICM 4 N.C R.B (N.C.R.B). I.F.- (Unified Investigation Form-l) Register the case according to the rules. Signature illegible, Thanadhiksh Officer 16.2.2024`	`Where does the petitioner Manju Yadav reside?`

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: 25
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: 25
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

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}
}

RK-1235
/

bge-base-FIR-matryoshka-BASELINE-10ep