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 |
|---|---|
| 6.0 |
|
| 4.0 |
|
| 9.0 |
|
| 13.0 |
|
| 7.0 |
|
| 11.0 |
|
| 3.0 |
|
| 12.0 |
|
| 0.0 |
|
| 5.0 |
|
| 1.0 |
|
| 2.0 |
|
| 10.0 |
|
| 8.0 |
|
| 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_fi0")
# Run inference
preds = model("다용도 방수 알미늄 시트지 방유 주방 가구/인테리어>DIY자재/용품>시트지>타일시트지")
| Training set | Min | Median | Max |
|---|---|---|---|
| Word count | 2 | 9.0153 | 20 |
| 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 |
| 8.0 | 70 |
| 9.0 | 70 |
| 10.0 | 70 |
| 11.0 | 70 |
| 12.0 | 70 |
| 13.0 | 70 |
| Epoch | Step | Training Loss | Validation Loss |
|---|---|---|---|
| 0.0052 | 1 | 0.4943 | - |
| 0.2604 | 50 | 0.497 | - |
| 0.5208 | 100 | 0.4938 | - |
| 0.7812 | 150 | 0.454 | - |
| 1.0417 | 200 | 0.31 | - |
| 1.3021 | 250 | 0.0825 | - |
| 1.5625 | 300 | 0.0174 | - |
| 1.8229 | 350 | 0.0104 | - |
| 2.0833 | 400 | 0.0018 | - |
| 2.3438 | 450 | 0.0002 | - |
| 2.6042 | 500 | 0.0001 | - |
| 2.8646 | 550 | 0.0001 | - |
| 3.125 | 600 | 0.0001 | - |
| 3.3854 | 650 | 0.0001 | - |
| 3.6458 | 700 | 0.0001 | - |
| 3.9062 | 750 | 0.0 | - |
| 4.1667 | 800 | 0.0 | - |
| 4.4271 | 850 | 0.0 | - |
| 4.6875 | 900 | 0.0 | - |
| 4.9479 | 950 | 0.0 | - |
| 5.2083 | 1000 | 0.0 | - |
| 5.4688 | 1050 | 0.0 | - |
| 5.7292 | 1100 | 0.0 | - |
| 5.9896 | 1150 | 0.0 | - |
| 6.25 | 1200 | 0.0 | - |
| 6.5104 | 1250 | 0.0 | - |
| 6.7708 | 1300 | 0.0 | - |
| 7.0312 | 1350 | 0.0 | - |
| 7.2917 | 1400 | 0.0 | - |
| 7.5521 | 1450 | 0.0 | - |
| 7.8125 | 1500 | 0.0 | - |
| 8.0729 | 1550 | 0.0 | - |
| 8.3333 | 1600 | 0.0 | - |
| 8.5938 | 1650 | 0.0 | - |
| 8.8542 | 1700 | 0.0 | - |
| 9.1146 | 1750 | 0.0 | - |
| 9.375 | 1800 | 0.0 | - |
| 9.6354 | 1850 | 0.0 | - |
| 9.8958 | 1900 | 0.0 | - |
| 10.1562 | 1950 | 0.0 | - |
| 10.4167 | 2000 | 0.0 | - |
| 10.6771 | 2050 | 0.0 | - |
| 10.9375 | 2100 | 0.0 | - |
| 11.1979 | 2150 | 0.0 | - |
| 11.4583 | 2200 | 0.0 | - |
| 11.7188 | 2250 | 0.0 | - |
| 11.9792 | 2300 | 0.0 | - |
| 12.2396 | 2350 | 0.0 | - |
| 12.5 | 2400 | 0.0 | - |
| 12.7604 | 2450 | 0.0 | - |
| 13.0208 | 2500 | 0.0 | - |
| 13.2812 | 2550 | 0.0 | - |
| 13.5417 | 2600 | 0.0 | - |
| 13.8021 | 2650 | 0.0 | - |
| 14.0625 | 2700 | 0.0 | - |
| 14.3229 | 2750 | 0.0 | - |
| 14.5833 | 2800 | 0.0 | - |
| 14.8438 | 2850 | 0.0 | - |
| 15.1042 | 2900 | 0.0 | - |
| 15.3646 | 2950 | 0.0 | - |
| 15.625 | 3000 | 0.0 | - |
| 15.8854 | 3050 | 0.0 | - |
| 16.1458 | 3100 | 0.0 | - |
| 16.4062 | 3150 | 0.0 | - |
| 16.6667 | 3200 | 0.0 | - |
| 16.9271 | 3250 | 0.0 | - |
| 17.1875 | 3300 | 0.0 | - |
| 17.4479 | 3350 | 0.0 | - |
| 17.7083 | 3400 | 0.0 | - |
| 17.9688 | 3450 | 0.0 | - |
| 18.2292 | 3500 | 0.0 | - |
| 18.4896 | 3550 | 0.0 | - |
| 18.75 | 3600 | 0.0 | - |
| 19.0104 | 3650 | 0.0 | - |
| 19.2708 | 3700 | 0.0 | - |
| 19.5312 | 3750 | 0.0 | - |
| 19.7917 | 3800 | 0.0 | - |
| 20.0521 | 3850 | 0.0 | - |
| 20.3125 | 3900 | 0.0 | - |
| 20.5729 | 3950 | 0.0 | - |
| 20.8333 | 4000 | 0.0 | - |
| 21.0938 | 4050 | 0.0 | - |
| 21.3542 | 4100 | 0.0 | - |
| 21.6146 | 4150 | 0.0 | - |
| 21.875 | 4200 | 0.0 | - |
| 22.1354 | 4250 | 0.0 | - |
| 22.3958 | 4300 | 0.0 | - |
| 22.6562 | 4350 | 0.0 | - |
| 22.9167 | 4400 | 0.0 | - |
| 23.1771 | 4450 | 0.0 | - |
| 23.4375 | 4500 | 0.0 | - |
| 23.6979 | 4550 | 0.0 | - |
| 23.9583 | 4600 | 0.0 | - |
| 24.2188 | 4650 | 0.0 | - |
| 24.4792 | 4700 | 0.0 | - |
| 24.7396 | 4750 | 0.0 | - |
| 25.0 | 4800 | 0.0 | - |
| 25.2604 | 4850 | 0.0 | - |
| 25.5208 | 4900 | 0.0 | - |
| 25.7812 | 4950 | 0.0 | - |
| 26.0417 | 5000 | 0.0 | - |
| 26.3021 | 5050 | 0.0 | - |
| 26.5625 | 5100 | 0.0 | - |
| 26.8229 | 5150 | 0.0 | - |
| 27.0833 | 5200 | 0.0 | - |
| 27.3438 | 5250 | 0.0 | - |
| 27.6042 | 5300 | 0.0 | - |
| 27.8646 | 5350 | 0.0 | - |
| 28.125 | 5400 | 0.0 | - |
| 28.3854 | 5450 | 0.0 | - |
| 28.6458 | 5500 | 0.0 | - |
| 28.9062 | 5550 | 0.0 | - |
| 29.1667 | 5600 | 0.0 | - |
| 29.4271 | 5650 | 0.0 | - |
| 29.6875 | 5700 | 0.0 | - |
| 29.9479 | 5750 | 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}
}