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 |
|
| 1.0 |
|
| 10.0 |
|
| 7.0 |
|
| 4.0 |
|
| 9.0 |
|
| 0.0 |
|
| 8.0 |
|
| 2.0 |
|
| 3.0 |
|
| 5.0 |
|
| Label | Accuracy |
|---|---|
| all | 0.7989 |
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_bt8_test")
# Run inference
preds = model("참존 탑클래스 리프팅 스킨 120ml 옵션없음 하루뷰티")
| Training set | Min | Median | Max |
|---|---|---|---|
| Word count | 4 | 9.2179 | 23 |
| Label | Training Sample Count |
|---|---|
| 0.0 | 18 |
| 1.0 | 18 |
| 2.0 | 22 |
| 3.0 | 20 |
| 4.0 | 32 |
| 5.0 | 30 |
| 6.0 | 40 |
| 7.0 | 23 |
| 8.0 | 17 |
| 9.0 | 14 |
| 10.0 | 23 |
| Epoch | Step | Training Loss | Validation Loss |
|---|---|---|---|
| 0.0323 | 1 | 0.4874 | - |
| 1.6129 | 50 | 0.3751 | - |
| 3.2258 | 100 | 0.0862 | - |
| 4.8387 | 150 | 0.0251 | - |
| 6.4516 | 200 | 0.0101 | - |
| 8.0645 | 250 | 0.0042 | - |
| 9.6774 | 300 | 0.0045 | - |
| 11.2903 | 350 | 0.0044 | - |
| 12.9032 | 400 | 0.0041 | - |
| 14.5161 | 450 | 0.0043 | - |
| 16.1290 | 500 | 0.0042 | - |
| 17.7419 | 550 | 0.0042 | - |
| 19.3548 | 600 | 0.004 | - |
| 20.9677 | 650 | 0.0043 | - |
| 22.5806 | 700 | 0.0042 | - |
| 24.1935 | 750 | 0.004 | - |
| 25.8065 | 800 | 0.0004 | - |
| 27.4194 | 850 | 0.0001 | - |
| 29.0323 | 900 | 0.0001 | - |
| 30.6452 | 950 | 0.0001 | - |
| 32.2581 | 1000 | 0.0001 | - |
| 33.8710 | 1050 | 0.0001 | - |
| 35.4839 | 1100 | 0.0001 | - |
| 37.0968 | 1150 | 0.0001 | - |
| 38.7097 | 1200 | 0.0001 | - |
| 40.3226 | 1250 | 0.0001 | - |
| 41.9355 | 1300 | 0.0001 | - |
| 43.5484 | 1350 | 0.0001 | - |
| 45.1613 | 1400 | 0.0001 | - |
| 46.7742 | 1450 | 0.0001 | - |
| 48.3871 | 1500 | 0.0001 | - |
| 50.0 | 1550 | 0.0001 | - |
@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}
}