This is a SetFit model that can be used for Text Classification. This SetFit model uses sentence-transformers/paraphrase-multilingual-MiniLM-L12-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-multilingual-MiniLM-L12-v2
Classification head: a LogisticRegression instance
Maximum Sequence Length: 128 tokens
Number of Classes: 8 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
Lanche	'x-tudo completo' 'x-bacon artesanal' 'hamburguer duplo'
Japonesa	'barca de sushi' 'temaki de salmão' 'sashimi'
Brasileira	'feijoada completa' 'prato feito (pf)' 'marmita'
Pizza/Massa	'pizza de calabresa' 'pizza portuguesa' 'pizza marguerita'
Sobremesa	'petit gateau' 'bolo de chocolate' 'torta de limão'
Bebida	'coca-cola zero' 'guaraná' 'suco de laranja'
Petiscos	'batata frita' 'batata frita com queijo' 'porção de batata'
Árabe	'esfiha de carne' 'esfiha de queijo' 'esfiha de frango'

Uses

Direct Use for Inference

First install the SetFit library:

pip install setfit

Then you can load this model and run inference.

python

from setfit import SetFitModel

# Download from the 🤗 Hub
model = SetFitModel.from_pretrained("JoaoVitorr/ifood-classification-model-v5")
# Run inference
preds = model("mocotó")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	1	2.1222	5

Label	Training Sample Count
Bebida	27
Brasileira	27
Japonesa	23
Lanche	30
Petiscos	27
Pizza/Massa	21
Sobremesa	46
Árabe	20

Training Hyperparameters

batch_size: (32, 32)
num_epochs: (5, 5)
max_steps: -1
sampling_strategy: oversampling
num_iterations: 20
body_learning_rate: (1e-05, 1e-05)
head_learning_rate: 0.01
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
0.0036	1	0.2488	-
0.1805	50	0.257	-
0.3610	100	0.2371	-
0.5415	150	0.2231	-
0.7220	200	0.198	-
0.9025	250	0.1617	-
1.0830	300	0.1286	-
1.2635	350	0.1051	-
1.4440	400	0.0908	-
1.6245	450	0.0757	-
1.8051	500	0.0619	-
1.9856	550	0.0465	-
2.1661	600	0.0355	-
2.3466	650	0.0304	-
2.5271	700	0.0218	-
2.7076	750	0.018	-
2.8881	800	0.0144	-
3.0686	850	0.0119	-
3.2491	900	0.0106	-
3.4296	950	0.008	-
3.6101	1000	0.0089	-
3.7906	1050	0.0083	-
3.9711	1100	0.0073	-
4.1516	1150	0.006	-
4.3321	1200	0.0058	-
4.5126	1250	0.0053	-
4.6931	1300	0.0052	-
4.8736	1350	0.0046	-

Framework Versions

Python: 3.12.12
SetFit: 1.1.3
Sentence Transformers: 5.1.2
Transformers: 4.57.1
PyTorch: 2.8.0+cu126
Datasets: 4.0.0
Tokenizers: 0.22.1

Citation

BibTeX

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}
}