JohanHeinsen/Labour_ads_male

SetFit with JohanHeinsen/Old_News_Segmentation_SBERT_V0.1

This is a SetFit model that can be used for Text Classification. This SetFit model uses JohanHeinsen/Old_News_Segmentation_SBERT_V0.1 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: JohanHeinsen/Old_News_Segmentation_SBERT_V0.1
• 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
0	'En skikkelig Pige søger Condition strar eller til St. Hansdag som Opvartningspige, i Mangel deraf som Stue= eller Enevighvor Konen gaaer i Huusholdningen, anvises paa Hiørnet af Larsbjørnstræde og Volden 236 i Stuen.' 'En Pige fra Landet søger strax Condition for Amme eller Goldamme, er at finde paa Vesterbro Nr. 9.' 'En Kone, der godt kan vaske, stryge og tillige godt lave Mad, ønsker sig Condition hos en honet Familie som Kokke eller Enepige, eller og at gaae i ugeviis, hun kan tillige i malke, om forlanges, anvises i Bredgaden Nr. 202 paa 5 første Sal.'
1	'En skikkelig Karl fra Jylland søger Condition til St. Hansdag og er at finde paa Christianshavn paa Hiørnet af Dronningensgade og Torvegagen i Kielderen i Nr. 359.' 'En svensk Karl, nyelig kommen her til Staden, ønsker sig Condition som Kudsk eller Tienercher i Byen, eller paa Landet, har sine behørige Skudsmaal, er at finde i store Kongensgade No. 51.' 'En Student, der er øvet i at informere, tilbyder sig at give Underviisning i det tydske Sprog, Regning, Skrivning, Religion, samt andre til Akademiet hørende Videnskaber Anviisningen gives i Adelgaden Nr. 206, første Sal, det første Huus paa høire Haand fra Gottersgaden.'

Evaluation

Metrics

Label	Accuracy	F1	Precision	Recall
all	0.9924	0.9882	0.9882	0.9882

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("En Stuepige, som forstaaer hvad hun bør, søger til Paaske; er at finde i Dronningens Tvergade Nr. 363 i Stuen.")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	8	32.4388	176

Label	Training Sample Count
0	419
1	194

Training Hyperparameters

• batch_size: (16, 16)
• num_epochs: (2, 2)
• max_steps: -1
• sampling_strategy: oversampling
• num_iterations: 12
• 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
0.0011	1	0.2907	-
0.0543	50	0.2347	-
0.1087	100	0.032	-
0.1630	150	0.0129	-
0.2174	200	0.0015	-
0.2717	250	0.0004	-
0.3261	300	0.0002	-
0.3804	350	0.0002	-
0.4348	400	0.0001	-
0.4891	450	0.0001	-
0.5435	500	0.0001	-
0.5978	550	0.0001	-
0.6522	600	0.0001	-
0.7065	650	0.0001	-
0.7609	700	0.0001	-
0.8152	750	0.0001	-
0.8696	800	0.0001	-
0.9239	850	0.0	-
0.9783	900	0.0	-
1.0326	950	0.0	-
1.0870	1000	0.0	-
1.1413	1050	0.0	-
1.1957	1100	0.0	-
1.25	1150	0.0	-
1.3043	1200	0.0	-
1.3587	1250	0.0	-
1.4130	1300	0.0	-
1.4674	1350	0.0	-
1.5217	1400	0.0	-
1.5761	1450	0.0	-
1.6304	1500	0.0	-
1.6848	1550	0.0	-
1.7391	1600	0.0	-
1.7935	1650	0.0	-
1.8478	1700	0.0	-
1.9022	1750	0.0	-
1.9565	1800	0.0	-

Framework Versions

• Python: 3.11.12
• SetFit: 1.1.3
• Sentence Transformers: 4.1.0
• Transformers: 4.51.3
• PyTorch: 2.7.0
• Datasets: 2.19.2
• 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}
}