Efficient Few-Shot Learning Without Prompts
Paper
•
2209.11055
•
Published
•
4
This is a SetFit model trained on the afk-live/afk-news-fr-classification-20251225 dataset that can be used for Text Classification. This SetFit model uses OrdalieTech/Solon-embeddings-large-0.1 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 |
|---|---|
| Sport |
|
| Lifestyle & Consumer (mode, cuisine, gadgets, pub) |
|
| Politique Europe & Francophonie (BE, CH, QC + UE) |
|
| Technologie |
|
| Climat & Biodiversité |
|
| Géopolitique mondiale (conflits, diplomatie hors Europe) |
|
| Faits divers & Actualité locale |
|
| Politique France |
|
| Science, Santé & Recherche |
|
| Culture |
|
| Outlier |
|
| Économie & Finance |
|
| Météo |
|
| Médias & People |
|
| Société & Éthique |
|
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("afk-live/afk-setfit-solon-large-v1")
# Run inference
preds = model("Bretagne : une maison rasée et des prairies bétonnées")
| Training set | Min | Median | Max |
|---|---|---|---|
| Word count | 1 | 13.6133 | 27 |
| Label | Training Sample Count |
|---|---|
| Climat & Biodiversité | 40 |
| Culture | 40 |
| Faits divers & Actualité locale | 40 |
| Géopolitique mondiale (conflits, diplomatie hors Europe) | 40 |
| Lifestyle & Consumer (mode, cuisine, gadgets, pub) | 40 |
| Médias & People | 40 |
| Météo | 40 |
| Outlier | 40 |
| Politique Europe & Francophonie (BE, CH, QC + UE) | 40 |
| Politique France | 40 |
| Science, Santé & Recherche | 40 |
| Société & Éthique | 40 |
| Sport | 40 |
| Technologie | 40 |
| Économie & Finance | 40 |
| Epoch | Step | Training Loss | Validation Loss |
|---|---|---|---|
| 0.0002 | 1 | 0.0662 | - |
| 0.0089 | 50 | 0.0674 | - |
| 0.0177 | 100 | 0.0607 | - |
| 0.0266 | 150 | 0.0556 | - |
| 0.0355 | 200 | 0.0567 | - |
| 0.0444 | 250 | 0.0464 | - |
| 0.0532 | 300 | 0.0476 | - |
| 0.0621 | 350 | 0.0484 | - |
| 0.0710 | 400 | 0.0369 | - |
| 0.0799 | 450 | 0.0327 | - |
| 0.0887 | 500 | 0.034 | - |
| 0.0976 | 550 | 0.031 | - |
| 0.1065 | 600 | 0.0257 | - |
| 0.1154 | 650 | 0.024 | - |
| 0.1242 | 700 | 0.0165 | - |
| 0.1331 | 750 | 0.0136 | - |
| 0.1420 | 800 | 0.0117 | - |
| 0.1508 | 850 | 0.0066 | - |
| 0.1597 | 900 | 0.0111 | - |
| 0.1686 | 950 | 0.0083 | - |
| 0.1775 | 1000 | 0.0058 | - |
| 0.1863 | 1050 | 0.0035 | - |
| 0.1952 | 1100 | 0.0023 | - |
| 0.2041 | 1150 | 0.0035 | - |
| 0.2130 | 1200 | 0.0034 | - |
| 0.2218 | 1250 | 0.0016 | - |
| 0.2307 | 1300 | 0.0007 | - |
| 0.2396 | 1350 | 0.0022 | - |
| 0.2484 | 1400 | 0.0031 | - |
| 0.2573 | 1450 | 0.0015 | - |
| 0.2662 | 1500 | 0.0016 | - |
| 0.2751 | 1550 | 0.0016 | - |
| 0.2839 | 1600 | 0.0026 | - |
| 0.2928 | 1650 | 0.001 | - |
| 0.3017 | 1700 | 0.0019 | - |
| 0.3106 | 1750 | 0.001 | - |
| 0.3194 | 1800 | 0.0013 | - |
| 0.3283 | 1850 | 0.0009 | - |
| 0.3372 | 1900 | 0.0008 | - |
| 0.3461 | 1950 | 0.0017 | - |
| 0.3549 | 2000 | 0.0018 | - |
| 0.3638 | 2050 | 0.0005 | - |
| 0.3727 | 2100 | 0.0005 | - |
| 0.3815 | 2150 | 0.0005 | - |
| 0.3904 | 2200 | 0.001 | - |
| 0.3993 | 2250 | 0.0002 | - |
| 0.4082 | 2300 | 0.0017 | - |
| 0.4170 | 2350 | 0.0005 | - |
| 0.4259 | 2400 | 0.0005 | - |
| 0.4348 | 2450 | 0.0002 | - |
| 0.4437 | 2500 | 0.0001 | - |
| 0.4525 | 2550 | 0.0001 | - |
| 0.4614 | 2600 | 0.0001 | - |
| 0.4703 | 2650 | 0.0001 | - |
| 0.4791 | 2700 | 0.0001 | - |
| 0.4880 | 2750 | 0.0001 | - |
| 0.4969 | 2800 | 0.0001 | - |
| 0.5058 | 2850 | 0.0001 | - |
| 0.5146 | 2900 | 0.0001 | - |
| 0.5235 | 2950 | 0.0001 | - |
| 0.5324 | 3000 | 0.0001 | - |
| 0.5413 | 3050 | 0.0001 | - |
| 0.5501 | 3100 | 0.0001 | - |
| 0.5590 | 3150 | 0.0001 | - |
| 0.5679 | 3200 | 0.0001 | - |
| 0.5768 | 3250 | 0.0001 | - |
| 0.5856 | 3300 | 0.0001 | - |
| 0.5945 | 3350 | 0.0001 | - |
| 0.6034 | 3400 | 0.0001 | - |
| 0.6122 | 3450 | 0.0001 | - |
| 0.6211 | 3500 | 0.0001 | - |
| 0.6300 | 3550 | 0.0001 | - |
| 0.6389 | 3600 | 0.0001 | - |
| 0.6477 | 3650 | 0.0001 | - |
| 0.6566 | 3700 | 0.0001 | - |
| 0.6655 | 3750 | 0.0001 | - |
| 0.6744 | 3800 | 0.0001 | - |
| 0.6832 | 3850 | 0.0001 | - |
| 0.6921 | 3900 | 0.0001 | - |
| 0.7010 | 3950 | 0.0001 | - |
| 0.7098 | 4000 | 0.0001 | - |
| 0.7187 | 4050 | 0.0001 | - |
| 0.7276 | 4100 | 0.0001 | - |
| 0.7365 | 4150 | 0.0001 | - |
| 0.7453 | 4200 | 0.0001 | - |
| 0.7542 | 4250 | 0.0001 | - |
| 0.7631 | 4300 | 0.0001 | - |
| 0.7720 | 4350 | 0.0001 | - |
| 0.7808 | 4400 | 0.0001 | - |
| 0.7897 | 4450 | 0.0001 | - |
| 0.7986 | 4500 | 0.0001 | - |
| 0.8075 | 4550 | 0.0001 | - |
| 0.8163 | 4600 | 0.0001 | - |
| 0.8252 | 4650 | 0.0001 | - |
| 0.8341 | 4700 | 0.0001 | - |
| 0.8429 | 4750 | 0.0001 | - |
| 0.8518 | 4800 | 0.0001 | - |
| 0.8607 | 4850 | 0.0001 | - |
| 0.8696 | 4900 | 0.0001 | - |
| 0.8784 | 4950 | 0.0001 | - |
| 0.8873 | 5000 | 0.0001 | - |
| 0.8962 | 5050 | 0.0001 | - |
| 0.9051 | 5100 | 0.0001 | - |
| 0.9139 | 5150 | 0.0001 | - |
| 0.9228 | 5200 | 0.0001 | - |
| 0.9317 | 5250 | 0.0001 | - |
| 0.9406 | 5300 | 0.0001 | - |
| 0.9494 | 5350 | 0.0001 | - |
| 0.9583 | 5400 | 0.0001 | - |
| 0.9672 | 5450 | 0.0001 | - |
| 0.9760 | 5500 | 0.0001 | - |
| 0.9849 | 5550 | 0.0001 | - |
| 0.9938 | 5600 | 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}
}
Base model
OrdalieTech/Solon-embeddings-large-0.1