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 |
|
| 5.0 |
|
| 2.0 |
|
| 8.0 |
|
| 4.0 |
|
| 7.0 |
|
| 0.0 |
|
| 3.0 |
|
| Label | Metric |
|---|---|
| all | 0.9102 |
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_fd9")
# Run inference
preds = model("본죽 쇠고기 장조림 170g x 4 마이엘(Maiel)")
| Training set | Min | Median | Max |
|---|---|---|---|
| Word count | 3 | 10.1981 | 21 |
| Label | Training Sample Count |
|---|---|
| 0.0 | 50 |
| 1.0 | 42 |
| 2.0 | 22 |
| 3.0 | 50 |
| 4.0 | 50 |
| 5.0 | 50 |
| 6.0 | 50 |
| 7.0 | 50 |
| 8.0 | 50 |
| Epoch | Step | Training Loss | Validation Loss |
|---|---|---|---|
| 0.0154 | 1 | 0.4845 | - |
| 0.7692 | 50 | 0.2975 | - |
| 1.5385 | 100 | 0.0992 | - |
| 2.3077 | 150 | 0.0418 | - |
| 3.0769 | 200 | 0.0246 | - |
| 3.8462 | 250 | 0.0358 | - |
| 4.6154 | 300 | 0.0185 | - |
| 5.3846 | 350 | 0.0123 | - |
| 6.1538 | 400 | 0.0121 | - |
| 6.9231 | 450 | 0.0008 | - |
| 7.6923 | 500 | 0.0003 | - |
| 8.4615 | 550 | 0.0002 | - |
| 9.2308 | 600 | 0.0001 | - |
| 10.0 | 650 | 0.0001 | - |
| 10.7692 | 700 | 0.0001 | - |
| 11.5385 | 750 | 0.0002 | - |
| 12.3077 | 800 | 0.0001 | - |
| 13.0769 | 850 | 0.0001 | - |
| 13.8462 | 900 | 0.0001 | - |
| 14.6154 | 950 | 0.0001 | - |
| 15.3846 | 1000 | 0.0001 | - |
| 16.1538 | 1050 | 0.0001 | - |
| 16.9231 | 1100 | 0.0001 | - |
| 17.6923 | 1150 | 0.0001 | - |
| 18.4615 | 1200 | 0.0001 | - |
| 19.2308 | 1250 | 0.0001 | - |
| 20.0 | 1300 | 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}
}