Kevinger/setfit-hub-report

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:

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-mpnet-base-v2
• Classification head: a LogisticRegression instance
• Maximum Sequence Length: 512 tokens
• Number of Classes: 9 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
3	'A Reinvented True Detective Plays It Cool' "It's owl season in Massachusetts. Here's how to spot them" 'Taylor Swift class at Harvard: Professor needs to hire more teaching assistants'
6	'Springfield Mayor Domenic Sarno tests positive for COVID-19' 'How to Take Care of Your Skin in the Fall and Winter' 'Subbing plant-based milk for dairy options is a healthy decision'
2	'Mattel Has a New Cherokee Barbie. Not Everyone Is Happy About It.' 'Who Is Alan Garber, Harvards Interim President?' 'Springfield Marine training in Japan near Mount Fuji (Photos)'
0	'Heres which Northampton businesses might soon get all-alcohol liquor licenses' 'People in Business: Jan. 15, 2024' 'Come Home With Memories, Not a Shocking Phone Bill'
7	'3 Patriots vs. Chiefs predictions' 'Tuskegee vs. Alabama State How to watch college football' 'WMass Boys Basketball Season Stats Leaders: Who leads the region by class?'
8	'Biting Cold Sweeping U.S. Hits the South With an Unfamiliar Freeze' 'Some Sunday storms and sun - Boston News, Weather, Sports' 'More snow on the way in Mass. on Tuesday with slippery evening commute'
4	'title' 'This sentence is label' 'This sentence is label'
1	'Two cars crash through former Boston Market in Saugus' 'U.S. Naval Officer Who Helped China Is Sentenced to 2 Years in Prison' 'American Airlines flight attendant arrested after allegedly filming teenage girl in bathroom on flight to Boston - Boston News, Weather, Sports'
5	'Opinion

Evaluation

Metrics

Label	Accuracy
all	0.7061

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("Kevinger/setfit-hub-report")
# Run inference
preds = model("Opinion | When the World Feels Dark, Seek Out Delight")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	1	7.2993	21

Label	Training Sample Count
0	16
1	16
2	16
3	16
4	9
5	16
6	16
7	16
8	16

Training Hyperparameters

• batch_size: (16, 2)
• num_epochs: (1, 16)
• 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
• seed: 42
• eval_max_steps: -1
• load_best_model_at_end: False

Training Results

Epoch	Step	Training Loss	Validation Loss
0.0010	1	0.3619	-
0.0481	50	0.097	-
0.0962	100	0.0327	-
0.1442	150	0.0044	-
0.1923	200	0.0013	-
0.2404	250	0.0011	-
0.2885	300	0.001	-
0.3365	350	0.0008	-
0.3846	400	0.001	-
0.4327	450	0.0006	-
0.4808	500	0.0008	-
0.5288	550	0.0005	-
0.5769	600	0.0012	-
0.625	650	0.0005	-
0.6731	700	0.0006	-
0.7212	750	0.0004	-
0.7692	800	0.0005	-
0.8173	850	0.0005	-
0.8654	900	0.0006	-
0.9135	950	0.0014	-
0.9615	1000	0.0006	-

Framework Versions

• Python: 3.10.12
• SetFit: 1.0.3
• Sentence Transformers: 2.2.2
• Transformers: 4.35.2
• PyTorch: 2.1.0+cu121
• Datasets: 2.16.1
• Tokenizers: 0.15.0

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}
}