gehaustein/PolyQual-3

SetFit with sentence-transformers/all-mpnet-base-v2

This is a SetFit model that can be used for Text Classification. This SetFit model uses sentence-transformers/all-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/all-mpnet-base-v2
• Classification head: a LogisticRegression instance
• Maximum Sequence Length: 384 tokens
• Number of Classes: 3 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	'maybe coffeezilla was right that Polymarket is just a tool for insiders to make money....' 'Coping yes kids in chat ' 'The most common with ABOP'
1	'Right now, I am just waiting for the shares for march 31st to go back down to 26-31c so I can buy more shares with what I won so far' 'If I am insider, I would do the opposite, scoop up yes quietly and not tell influencers. By the time you see it tweeter you are the liquidity' 'Looks like Syrian government bet NO in this market :) https://kyivindependent.com/russias-evacuation-efforts-stalled-as-new-syrian-leaders-deny-port-access-media-reports/'
2	'Somewhere in GOP 1-64 is where I think it'll end up. Feel good about PA, GA, and NC. Normally Trump underpolls big time in Wisconsin, but recent statewide elections there don't look good for Trump. Michigan was a fluke in 2016, when Dems are focused that state is so hard to flip. AZ is full of McCain & Flake "cuck" Republicans so don't feel good there about Trump's chances, NV is a better flip opportunity in the Southwest imho.' 'According to multiple sources, President Joe Biden signed a bill to avoid a government shutdown on December 20, 2024. This action was reported across various news outlets and social media: The Senate passed a stopgap funding bill shortly after the midnight funding deadline, and the House had passed it earlier that evening. President Biden was set to sign this legislation, which would extend government funding into March and include provisions for disaster relief and farm aid. Posts on X also confirmed that the bill was passed by Congress and sent to President Biden for signing, with the explicit mention that it averted a shutdown. These sources collectively indicate that the signing took place on December 20, 2024.' "Looks like he's about to do it https://apnews.com/article/trump-deportees-el-salvador-contempt-boasberg-da282511ac6f5c8dd19af620995ca440"

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("gehaustein/PolyQual-3")
# Run inference
preds = model("Lol prove it")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	1	22.3420	199

Label	Training Sample Count
0	307
1	307
2	307

Training Hyperparameters

• batch_size: (8, 8)
• num_epochs: (1, 1)
• max_steps: -1
• sampling_strategy: oversampling
• num_iterations: 20
• 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
0.0002	1	0.6034	-
0.0109	50	0.3441	0.3746
0.0217	100	0.3198	0.3002
0.0326	150	0.2498	0.2823
0.0434	200	0.2468	0.2755
0.0543	250	0.2242	0.2678
0.0651	300	0.174	0.2492
0.0760	350	0.1182	0.2157
0.0869	400	0.0824	0.2100
0.0977	450	0.0433	0.2346
0.1086	500	0.0248	0.2168
0.1194	550	0.0183	0.2211

Framework Versions

• Python: 3.12.13
• SetFit: 1.1.3
• Sentence Transformers: 5.3.0
• Transformers: 4.49.0
• PyTorch: 2.10.0+cu128
• Datasets: 4.0.0
• Tokenizers: 0.21.4

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}
}