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 |
|---|---|
| 0.0 |
|
| 4.0 |
|
| 2.0 |
|
| 5.0 |
|
| 3.0 |
|
| 1.0 |
|
| 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_bc21")
# Run inference
preds = model("[국제금거래소] (순도99.9%) 고급 순금 돌반지 1.875g 복(福)_고급케이스 출산/육아 > 유아동주얼리 > 순금돌반지")
| Training set | Min | Median | Max |
|---|---|---|---|
| Word count | 7 | 15.7703 | 32 |
| Label | Training Sample Count |
|---|---|
| 0.0 | 70 |
| 1.0 | 20 |
| 2.0 | 70 |
| 3.0 | 70 |
| 4.0 | 70 |
| 5.0 | 70 |
| Epoch | Step | Training Loss | Validation Loss |
|---|---|---|---|
| 0.0137 | 1 | 0.4867 | - |
| 0.6849 | 50 | 0.4987 | - |
| 1.3699 | 100 | 0.3808 | - |
| 2.0548 | 150 | 0.1425 | - |
| 2.7397 | 200 | 0.053 | - |
| 3.4247 | 250 | 0.0037 | - |
| 4.1096 | 300 | 0.0001 | - |
| 4.7945 | 350 | 0.0001 | - |
| 5.4795 | 400 | 0.0001 | - |
| 6.1644 | 450 | 0.0001 | - |
| 6.8493 | 500 | 0.0 | - |
| 7.5342 | 550 | 0.0 | - |
| 8.2192 | 600 | 0.0 | - |
| 8.9041 | 650 | 0.0 | - |
| 9.5890 | 700 | 0.0 | - |
| 10.2740 | 750 | 0.0 | - |
| 10.9589 | 800 | 0.0 | - |
| 11.6438 | 850 | 0.0 | - |
| 12.3288 | 900 | 0.0 | - |
| 13.0137 | 950 | 0.0 | - |
| 13.6986 | 1000 | 0.0 | - |
| 14.3836 | 1050 | 0.0 | - |
| 15.0685 | 1100 | 0.0 | - |
| 15.7534 | 1150 | 0.0 | - |
| 16.4384 | 1200 | 0.0 | - |
| 17.1233 | 1250 | 0.0 | - |
| 17.8082 | 1300 | 0.0 | - |
| 18.4932 | 1350 | 0.0 | - |
| 19.1781 | 1400 | 0.0 | - |
| 19.8630 | 1450 | 0.0 | - |
| 20.5479 | 1500 | 0.0 | - |
| 21.2329 | 1550 | 0.0 | - |
| 21.9178 | 1600 | 0.0 | - |
| 22.6027 | 1650 | 0.0 | - |
| 23.2877 | 1700 | 0.0 | - |
| 23.9726 | 1750 | 0.0 | - |
| 24.6575 | 1800 | 0.0 | - |
| 25.3425 | 1850 | 0.0 | - |
| 26.0274 | 1900 | 0.0 | - |
| 26.7123 | 1950 | 0.0 | - |
| 27.3973 | 2000 | 0.0 | - |
| 28.0822 | 2050 | 0.0 | - |
| 28.7671 | 2100 | 0.0 | - |
| 29.4521 | 2150 | 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}
}