faodl/v02_model_child_and_family_support_benefits_mpnet_60_sample

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

This is a SetFit model that can be used for Text Classification. This SetFit model uses sentence-transformers/paraphrase-multilingual-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:

1. Fine-tuning a Sentence Transformer with contrastive learning.
2. 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-mpnet-base-v2
• Classification head: a LogisticRegression instance
• Maximum Sequence Length: 128 tokens
• Number of Classes: 2 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
Relevant	'Complementing cash transfers with nutrition and health services exemplifies a comprehensive approach that addresses the multifaceted needs of children in impoverished communities.' 'Non-contributory support mechanisms, financed through taxation, are vital to reaching children and families excluded from formal employment-based schemes, thus addressing structural inequalities.' 'Predictable disbursement schedules foster trust and reliability, enabling caregivers to secure consistent access to essential resources for their children.'
Irrelevant	'For maximum impact on nutrition, staff \nand volunteers will need clear and detailed guidelines on the content and priorities of outreach \nvisits, communities will need to be aware of the timing and place of the visits, and staff and \nvolunteers will need to keep accurate records of community members that should be present \nor visited during outreach (e.g.' '287Human Development, Poverty and Public Programmes\n\nBihar.' 'The\t\r \xa0 need\t\r \xa0 for\t\r \xa0 synchronized\t\r \xa0 and\t\r \xa0 automatic\t\r \xa0 weather\t\r \xa0 collection\t\r \xa0 systems\t\r \xa0 across\t\r \xa0 the\t\r \xa0 different\t\r \xa0 agro-\xad‐\necological\t\r \xa0zones\t\r \xa0of\t\r \xa0the\t\r \xa0country\t\r \xa0guarantees\t\r \xa0a\t\r \xa0higher\t\r \xa0data\t\r \xa0resolution\t\r \xa0for\t\r \xa0reliable\t\r \xa0data\t\r \xa0processing\t\r \xa0and\t\r \xa0\nallows\t\r \xa0 for\t\r \xa0 a\t\r \xa0 systematic\t\r \xa0 presentation\t\r \xa0 of\t\r \xa0 spatio-\xad‐temporal\t\r \xa0 weather\t\r \xa0 variability\t\r \xa0 and\t\r \xa0 mapping\t\r \xa0 of\t\r \xa0\nvulnerable\t\r \xa0areas\t\r \xa0(BNRCC,\t\r \xa02011).'

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("faodl/v02_model_child_and_family_support_benefits_mpnet_60_sample")
# Run inference
preds = model("STRATGEY FOR AGRICULTURE AND WATER – HARMONIZED PROGRAM DESIGN DOCUMENT – FINAL 

 

17 
 

3.4 Agriculture and Agribusiness  

 
107.")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	2	27.8438	180

Label	Training Sample Count
Irrelevant	48
Relevant	48

Training Hyperparameters

• batch_size: (16, 16)
• num_epochs: (1, 1)
• 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
0.0042	1	0.23	-
0.2083	50	0.1041	-
0.4167	100	0.0018	-
0.625	150	0.0006	-
0.8333	200	0.0004	-

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: 3.6.0
• 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}
}