Efficient Few-Shot Learning Without Prompts
Paper • 2209.11055 • Published • 7
How to use mini1013/master_cate_ac1 with setfit:
from setfit import SetFitModel
model = SetFitModel.from_pretrained("mini1013/master_cate_ac1")How to use mini1013/master_cate_ac1 with sentence-transformers:
from sentence_transformers import SentenceTransformer
model = SentenceTransformer("mini1013/master_cate_ac1")
sentences = [
"The weather is lovely today.",
"It's so sunny outside!",
"He drove to the stadium."
]
embeddings = model.encode(sentences)
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]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 |
|---|---|
| 10.0 |
|
| 4.0 |
|
| 7.0 |
|
| 3.0 | |
| 1.0 |
|
| 9.0 |
|
| 0.0 |
|
| 2.0 |
|
| 6.0 |
|
| 8.0 |
|
| 12.0 |
|
| 11.0 |
|
| 5.0 |
|
| Label | Metric |
|---|---|
| all | 0.5946 |
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_ac1")
# Run inference
preds = model("[프로스펙스 본사] 파워소닉 513 260 (주)엘에스네트웍스")
| Training set | Min | Median | Max |
|---|---|---|---|
| Word count | 3 | 10.5062 | 24 |
| Label | Training Sample Count |
|---|---|
| 0.0 | 50 |
| 1.0 | 50 |
| 2.0 | 50 |
| 3.0 | 50 |
| 4.0 | 50 |
| 5.0 | 50 |
| 6.0 | 50 |
| 7.0 | 50 |
| 8.0 | 50 |
| 9.0 | 50 |
| 10.0 | 50 |
| 11.0 | 50 |
| 12.0 | 50 |
| Epoch | Step | Training Loss | Validation Loss |
|---|---|---|---|
| 0.0098 | 1 | 0.4275 | - |
| 0.4902 | 50 | 0.3352 | - |
| 0.9804 | 100 | 0.2575 | - |
| 1.4706 | 150 | 0.1047 | - |
| 1.9608 | 200 | 0.0551 | - |
| 2.4510 | 250 | 0.0236 | - |
| 2.9412 | 300 | 0.0234 | - |
| 3.4314 | 350 | 0.0063 | - |
| 3.9216 | 400 | 0.0041 | - |
| 4.4118 | 450 | 0.0058 | - |
| 4.9020 | 500 | 0.0015 | - |
| 5.3922 | 550 | 0.0005 | - |
| 5.8824 | 600 | 0.0002 | - |
| 6.3725 | 650 | 0.0002 | - |
| 6.8627 | 700 | 0.0002 | - |
| 7.3529 | 750 | 0.0002 | - |
| 7.8431 | 800 | 0.0001 | - |
| 8.3333 | 850 | 0.0001 | - |
| 8.8235 | 900 | 0.0001 | - |
| 9.3137 | 950 | 0.0001 | - |
| 9.8039 | 1000 | 0.0001 | - |
| 10.2941 | 1050 | 0.0001 | - |
| 10.7843 | 1100 | 0.0001 | - |
| 11.2745 | 1150 | 0.0001 | - |
| 11.7647 | 1200 | 0.0001 | - |
| 12.2549 | 1250 | 0.0001 | - |
| 12.7451 | 1300 | 0.0001 | - |
| 13.2353 | 1350 | 0.0001 | - |
| 13.7255 | 1400 | 0.0001 | - |
| 14.2157 | 1450 | 0.0001 | - |
| 14.7059 | 1500 | 0.0001 | - |
| 15.1961 | 1550 | 0.0001 | - |
| 15.6863 | 1600 | 0.0001 | - |
| 16.1765 | 1650 | 0.0001 | - |
| 16.6667 | 1700 | 0.0001 | - |
| 17.1569 | 1750 | 0.0001 | - |
| 17.6471 | 1800 | 0.0001 | - |
| 18.1373 | 1850 | 0.0001 | - |
| 18.6275 | 1900 | 0.0001 | - |
| 19.1176 | 1950 | 0.0 | - |
| 19.6078 | 2000 | 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}
}