SetFit with sentence-transformers/paraphrase-mpnet-base-v2

This is a SetFit model that can be used for Text Classification. This SetFit model uses sentence-transformers/paraphrase-mpnet-base-v2 as the Sentence Transformer embedding model. A LogisticRegression instance is used for classification.

The model has been trained using an efficient few-shot learning technique that involves:

Fine-tuning a Sentence Transformer with contrastive learning.
Training a classification head with features from the fine-tuned Sentence Transformer.

Model Details

Model Description

Model Type: SetFit
Sentence Transformer body: sentence-transformers/paraphrase-mpnet-base-v2
Classification head: a LogisticRegression instance
Maximum Sequence Length: 512 tokens
Number of Classes: 10 classes

Model Sources

Repository: SetFit on GitHub
Paper: Efficient Few-Shot Learning Without Prompts
Blogpost: SetFit: Efficient Few-Shot Learning Without Prompts

Model Labels

Label	Examples
Groceries	'CARD NO.400536XXXXXX9172 LULU HYPERMARKET\ufffeWTC ABU DHABI:AE 312185 08-11-2024 0.75,AED' 'CARD NO.400536XXXXXX9172 NOON Minutes DUBAI:AE 467568 01-05-2025 25.00,AED' 'CARD NO.400536XXXXXX9172 LULU HYPERMARKET\ufffeWTC ABU DHABI:AE 128734 05-12-2024 5.60,AED'
Transport	'CARD NO.400536XXXXXX9172 Integrated Transport C Abu Dhabi:AE 123260 20-09-2024 20.00,AED' 'CARD NO.400536XXXXXX9172 Thrifty Rent A Car DUBAI:AE 654383 08-10-2024 147.00,AED' 'CARD NO.400536XXXXXX9172 CAREEM FOOD Dubai:AE 541167 17-11-2024 31.56,AED'
Shopping	'CARD NO.400536XXXXXX9172 MAJESTIC OPTICS COMPANY B ABU DHABI:AE 427849 25-09-2024 100.00,AED' 'CARD NO.400536XXXXXX9172 Temu.com Dublin 4:IE 566842 30-10-2024 25.58,AED' 'CARD NO.400536XXXXXX9172 AWS EMEA aws.amazon.co:LU 891744 02-10-2024 2.10,USD'
Transfer	'RULE TRANSFER TO KEVIN SAMSON WITH ONE-SHOT SAVING' 'MEPAY TRANSFER FROM SARANG SAJITH MURIKKO LI SAJIT H KATTINTAVIDE MURIKKOLI MOBILE NO. 009715XXXX2766 ; ; BNK REF.-80280-1927922' 'MEPAY TRANSFER TO ATM EI ATM -AIRPORT ROAD BRANCH MOBILE NO. 009715XXXX9454; ; BNK REF.-80280-169415 8'
Food	'CARD NO.400536XXXXXX9172 SMILES FOOD Abu Dhabi:AE 609846 05-01-2025 28.44,AED' 'CARD NO.400536XXXXXX9172 KFC Sharjah:AE 683418 15-11-2024 12.00,AED' 'CARD NO.400536XXXXXX9172 SASU CAFE Sharjah:AE 677254 05-11-2024 52.00,AED'
Salary	'IPP REF 20241226ADC6B9811094567470947 420168603 SM ART SOFTWARE SYSTEMS SOLUTIONS COMPUTER SERVICES' 'IPP REF 20240906ADC6B9811115537894689 374851134 SM ART SOFTWARE SYSTEMS SOLUTIONS COMPUTER SERVICES' 'IPP REF 20241127ADC6B9811155459103219 407878556 SM ART SOFTWARE SYSTEMS SOLUTIONS COMPUTER SERVICES'
Investment	'[email protected]' '[email protected]' '[email protected]'
Other	'TO BUY SOMETHING SPECIAL WITH SET AND FORGET' 'IPP REF 20240912E096B9811115139169618 56A5999A02A5 4D14A7580AE54 SHAHROZ MALIK MALIK MUHA' 'IPP REF 20240916NBA6B9811174440033095 FT2426063614 LOUEES IBRAHEM ALHANNA THANKS'
Utilities	'CARD NO.400536XXXXXX9172 ADNOC AL WIDAYAHI 166 ABUDHABI:AE 599558 05-11-2024 40.04,AED' 'CARD NO.400536XXXXXX9172 :27:33 E8001182 758692 AIRPORT ROAD BRANCH DUBAI AE' 'CARD NO.400536XXXXXX9172 BITS PILANI FZ LLC DUBAI:AE 395442 26-11-2024 300.00,AED'
Personal Care	'CARD NO.400536XXXXXX9172 CROWN RUBY SALON ABU DHABI:AE 101427 03-11-2024 25.00,AED' 'CARD NO.400536XXXXXX9172 CROWN RUBY SALON ABU DHABI:AE 270336 31-12-2024 25.00,AED'

Evaluation

Metrics

Label	Accuracy
all	0.8929

Uses

Direct Use for Inference

First install the SetFit library:

pip install setfit

Then you can load this model and run inference.

from setfit import SetFitModel

# Download from the 🤗 Hub
model = SetFitModel.from_pretrained("k3vin-samson/bank-transaction-classifeer")
# Run inference
preds = model("CARD NO.400536XXXXXX9172 CARS TAXI ABU DHABI:AE 882334 02-01-2025 14.50,AED")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	1	9.1696	18

Label	Training Sample Count
Food	23
Groceries	20
Investment	4
Other	8
Personal Care	2
Salary	3
Shopping	14
Transfer	21
Transport	12
Utilities	5

Training Hyperparameters

batch_size: (8, 8)
num_epochs: (3, 3)
max_steps: -1
sampling_strategy: oversampling
num_iterations: 20
body_learning_rate: (2e-05, 2e-05)
head_learning_rate: 2e-05
loss: CosineSimilarityLoss
distance_metric: cosine_distance
margin: 0.25
end_to_end: False
use_amp: False
warmup_proportion: 0.1
l2_weight: 0.01
seed: 42
eval_max_steps: -1
load_best_model_at_end: False

Training Results

Epoch	Step	Training Loss	Validation Loss
0.0018	1	0.2229	-
0.0893	50	0.1531	-
0.1786	100	0.0814	-
0.2679	150	0.0701	-
0.3571	200	0.0207	-
0.4464	250	0.0158	-
0.5357	300	0.0132	-
0.625	350	0.0055	-
0.7143	400	0.0026	-
0.8036	450	0.0013	-
0.8929	500	0.001	-
0.9821	550	0.0008	-
1.0714	600	0.0007	-
1.1607	650	0.0007	-
1.25	700	0.0004	-
1.3393	750	0.0004	-
1.4286	800	0.0003	-
1.5179	850	0.0003	-
1.6071	900	0.0005	-
1.6964	950	0.0005	-
1.7857	1000	0.0003	-
1.875	1050	0.0002	-
1.9643	1100	0.0003	-
2.0536	1150	0.0003	-
2.1429	1200	0.0002	-
2.2321	1250	0.0002	-
2.3214	1300	0.0003	-
2.4107	1350	0.0003	-
2.5	1400	0.0002	-
2.5893	1450	0.0002	-
2.6786	1500	0.0002	-
2.7679	1550	0.0002	-
2.8571	1600	0.0002	-
2.9464	1650	0.0002	-

Framework Versions

Python: 3.11.13
SetFit: 1.1.2
Sentence Transformers: 4.1.0
Transformers: 4.52.4
PyTorch: 2.6.0+cu124
Datasets: 2.14.4
Tokenizers: 0.21.1

Citation

BibTeX

@article{https://doi.org/10.48550/arxiv.2209.11055,
    doi = {10.48550/ARXIV.2209.11055},
    url = {https://arxiv.org/abs/2209.11055},
    author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren},
    keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences},
    title = {Efficient Few-Shot Learning Without Prompts},
    publisher = {arXiv},
    year = {2022},
    copyright = {Creative Commons Attribution 4.0 International}
}

k3vin-samson
/

setfit-bank-transaction-classifier