Efficient Few-Shot Learning Without Prompts
Paper
•
2209.11055
•
Published
•
4
This is a SetFit model that can be used for Text Classification. This SetFit model uses mini1013/master_domain 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:
| Label | Examples |
|---|---|
| 7.0 |
|
| 3.0 |
|
| 1.0 |
|
| 5.0 |
|
| 0.0 |
|
| 4.0 |
|
| 2.0 |
|
| 6.0 |
|
| Label | Accuracy |
|---|---|
| all | 1.0 |
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("mini1013/master_cate_bc26")
# Run inference
preds = model("4p 투데코 이유식 도자기 조리기세트 화이트 출산/육아 > 이유식용품 > 조리기")
| Training set | Min | Median | Max |
|---|---|---|---|
| Word count | 7 | 15.075 | 30 |
| Label | Training Sample Count |
|---|---|
| 0.0 | 70 |
| 1.0 | 70 |
| 2.0 | 70 |
| 3.0 | 70 |
| 4.0 | 70 |
| 5.0 | 70 |
| 6.0 | 70 |
| 7.0 | 70 |
| Epoch | Step | Training Loss | Validation Loss |
|---|---|---|---|
| 0.0091 | 1 | 0.4946 | - |
| 0.4545 | 50 | 0.5017 | - |
| 0.9091 | 100 | 0.4932 | - |
| 1.3636 | 150 | 0.3697 | - |
| 1.8182 | 200 | 0.0968 | - |
| 2.2727 | 250 | 0.0213 | - |
| 2.7273 | 300 | 0.0175 | - |
| 3.1818 | 350 | 0.0186 | - |
| 3.6364 | 400 | 0.0187 | - |
| 4.0909 | 450 | 0.0136 | - |
| 4.5455 | 500 | 0.0007 | - |
| 5.0 | 550 | 0.0001 | - |
| 5.4545 | 600 | 0.0001 | - |
| 5.9091 | 650 | 0.0001 | - |
| 6.3636 | 700 | 0.0001 | - |
| 6.8182 | 750 | 0.0001 | - |
| 7.2727 | 800 | 0.0001 | - |
| 7.7273 | 850 | 0.0001 | - |
| 8.1818 | 900 | 0.0 | - |
| 8.6364 | 950 | 0.0 | - |
| 9.0909 | 1000 | 0.0 | - |
| 9.5455 | 1050 | 0.0 | - |
| 10.0 | 1100 | 0.0 | - |
| 10.4545 | 1150 | 0.0 | - |
| 10.9091 | 1200 | 0.0 | - |
| 11.3636 | 1250 | 0.0 | - |
| 11.8182 | 1300 | 0.0 | - |
| 12.2727 | 1350 | 0.0 | - |
| 12.7273 | 1400 | 0.0 | - |
| 13.1818 | 1450 | 0.0 | - |
| 13.6364 | 1500 | 0.0 | - |
| 14.0909 | 1550 | 0.0 | - |
| 14.5455 | 1600 | 0.0 | - |
| 15.0 | 1650 | 0.0 | - |
| 15.4545 | 1700 | 0.0 | - |
| 15.9091 | 1750 | 0.0 | - |
| 16.3636 | 1800 | 0.0 | - |
| 16.8182 | 1850 | 0.0 | - |
| 17.2727 | 1900 | 0.0 | - |
| 17.7273 | 1950 | 0.0 | - |
| 18.1818 | 2000 | 0.0 | - |
| 18.6364 | 2050 | 0.0 | - |
| 19.0909 | 2100 | 0.0 | - |
| 19.5455 | 2150 | 0.0 | - |
| 20.0 | 2200 | 0.0 | - |
| 20.4545 | 2250 | 0.0 | - |
| 20.9091 | 2300 | 0.0 | - |
| 21.3636 | 2350 | 0.0 | - |
| 21.8182 | 2400 | 0.0 | - |
| 22.2727 | 2450 | 0.0 | - |
| 22.7273 | 2500 | 0.0 | - |
| 23.1818 | 2550 | 0.0 | - |
| 23.6364 | 2600 | 0.0 | - |
| 24.0909 | 2650 | 0.0 | - |
| 24.5455 | 2700 | 0.0 | - |
| 25.0 | 2750 | 0.0 | - |
| 25.4545 | 2800 | 0.0 | - |
| 25.9091 | 2850 | 0.0 | - |
| 26.3636 | 2900 | 0.0 | - |
| 26.8182 | 2950 | 0.0 | - |
| 27.2727 | 3000 | 0.0 | - |
| 27.7273 | 3050 | 0.0 | - |
| 28.1818 | 3100 | 0.0 | - |
| 28.6364 | 3150 | 0.0 | - |
| 29.0909 | 3200 | 0.0 | - |
| 29.5455 | 3250 | 0.0 | - |
| 30.0 | 3300 | 0.0 | - |
@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}
}