ychen/minor-persona-detection-en

SetFit with BAAI/bge-small-en-v1.5

This is a SetFit model that can be used for Text Classification. This SetFit model uses BAAI/bge-small-en-v1.5 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: BAAI/bge-small-en-v1.5
• Classification head: a LogisticRegression instance
• Maximum Sequence Length: 512 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
not_minor	'A young manager in her very first leadership position.' 'A detail-oriented student who excels in organizing group study sessions in the library' 'A fellow student involved in a book club who prefers physical copies for annotation and discussion'
minor	'A teenage girl from a disadvantaged background who is empowered by the health education programs' "A young child from a diverse family background who is involved in the candidate's research studies" 'A child with a passion for music who learns best through creative and interactive activities'

Evaluation

Metrics

Label	Accuracy
all	0.96

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("setfit_model_id")
# Run inference
preds = model("A cartoonist specializing in educational materials")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	3	14.34	26

Label	Training Sample Count
not_minor	100
minor	100

Training Hyperparameters

• batch_size: (64, 64)
• num_epochs: (4, 4)
• max_steps: -1
• sampling_strategy: oversampling
• body_learning_rate: (2e-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: True

Training Results

Epoch	Step	Training Loss	Validation Loss
0.0032	1	0.2636	-
0.1582	50	0.2471	-
0.3165	100	0.2067	-
0.4747	150	0.0207	-
0.6329	200	0.0021	-
0.7911	250	0.0015	-
0.9494	300	0.0013	-
1.0	316	-	0.0825
1.1076	350	0.0011	-
1.2658	400	0.001	-
1.4241	450	0.0009	-
1.5823	500	0.0008	-
1.7405	550	0.0008	-
1.8987	600	0.0007	-
2.0	632	-	0.0813
2.0570	650	0.001	-
2.2152	700	0.0007	-
2.3734	750	0.0007	-
2.5316	800	0.0006	-
2.6899	850	0.0006	-
2.8481	900	0.0006	-
3.0	948	-	0.0736
3.0063	950	0.0006	-
3.1646	1000	0.0006	-
3.3228	1050	0.0005	-
3.4810	1100	0.0006	-
3.6392	1150	0.0005	-
3.7975	1200	0.0006	-
3.9557	1250	0.0005	-
4.0	1264	-	0.0754

Framework Versions

• Python: 3.12.4
• SetFit: 1.1.0
• Sentence Transformers: 3.2.1
• Transformers: 4.45.2
• PyTorch: 2.5.1
• Datasets: 3.1.0
• Tokenizers: 0.20.3

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}
}