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 SetFitHead instance is used for classification.
The model has been trained using an efficient few-shot learning technique that involves:
- Fine-tuning a Sentence Transformer with contrastive learning.
- Training a classification head with features from the fine-tuned Sentence Transformer.
Model Details
Model Description
Model Sources
Model Labels
Label |
Examples |
8 |
- 'Later she meets someone at the bar. He'
- 'He heads to them and sits. The bus'
- 'Someone leaps to his feet and punches the agent in the face. Seemingly unaffected, the agent'
|
2 |
- 'A man sits behind a desk. Two people'
- 'A man is seen standing at the bottom of a hole while a man records him. Two men'
- 'Someone questions his female colleague who shrugs. Through a window, we'
|
0 |
- 'A woman bends down and puts something on a scale. She then'
- 'He pulls down the blind. He'
- 'Someone flings his hands forward. The someone fires, but the water'
|
6 |
- 'People are sitting down on chairs. They'
- 'They look up at stained glass skylights. The Americans'
- 'The lady and the man dance around each other in a circle. The people'
|
1 |
- 'An older gentleman kisses her. As he leads her off, someone'
- 'The first girl comes back and does it effortlessly as the second girl still struggles. For the last round, the girl'
- 'As she leaves, the bartender smiles. Now the blonde'
|
3 |
- 'Someone lowers his demoralized gaze. Someone'
- 'Someone goes into his bedroom. Someone'
- 'As someone leaves, someone spots him on the monitor. Someone'
|
7 |
- 'Four inches of Plexiglas separate the two and they talk on monitored phones. Someone'
- 'The American and Russian commanders each watch them returning. As someone'
- 'A group of walkers walk along the sidewalk near the lake. A man'
|
4 |
- 'The secretary flexes the foot of her crossed - leg as she eyes someone. The woman'
- 'A man in a white striped shirt is smiling. A woman'
- 'He grabs her hair and pulls her head back. She'
|
5 |
- 'He heads out of the plaza. Someone'
- "As he starts back, he sees someone's scared look just before he slams the door shut. Someone"
- 'He nods at her beaming. Someone'
|
Evaluation
Metrics
Label |
Accuracy |
all |
0.1654 |
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
model = SetFitModel.from_pretrained("HelgeKn/Swag-multi-class-20")
preds = model("He sneers and winds up with his fist. Someone")
Training Details
Training Set Metrics
Training set |
Min |
Median |
Max |
Word count |
5 |
12.1056 |
33 |
Label |
Training Sample Count |
0 |
20 |
1 |
20 |
2 |
20 |
3 |
20 |
4 |
20 |
5 |
20 |
6 |
20 |
7 |
20 |
8 |
20 |
Training Hyperparameters
- batch_size: (16, 16)
- num_epochs: (2, 2)
- 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
- seed: 42
- eval_max_steps: -1
- load_best_model_at_end: False
Training Results
Epoch |
Step |
Training Loss |
Validation Loss |
0.0022 |
1 |
0.3747 |
- |
0.1111 |
50 |
0.2052 |
- |
0.2222 |
100 |
0.1878 |
- |
0.3333 |
150 |
0.1126 |
- |
0.4444 |
200 |
0.1862 |
- |
0.5556 |
250 |
0.1385 |
- |
0.6667 |
300 |
0.0154 |
- |
0.7778 |
350 |
0.0735 |
- |
0.8889 |
400 |
0.0313 |
- |
1.0 |
450 |
0.0189 |
- |
1.1111 |
500 |
0.0138 |
- |
1.2222 |
550 |
0.0046 |
- |
1.3333 |
600 |
0.0043 |
- |
1.4444 |
650 |
0.0021 |
- |
1.5556 |
700 |
0.0033 |
- |
1.6667 |
750 |
0.001 |
- |
1.7778 |
800 |
0.0026 |
- |
1.8889 |
850 |
0.0022 |
- |
2.0 |
900 |
0.0014 |
- |
Framework Versions
- Python: 3.9.13
- SetFit: 1.0.1
- Sentence Transformers: 2.2.2
- Transformers: 4.36.0
- PyTorch: 2.1.1+cpu
- Datasets: 2.15.0
- 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}
}