SetFit with projecte-aina/ST-NLI-ca_paraphrase-multilingual-mpnet-base
This is a SetFit model that can be used for Text Classification. This SetFit model uses projecte-aina/ST-NLI-ca_paraphrase-multilingual-mpnet-base 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:
- Fine-tuning a Sentence Transformer with contrastive learning.
- Training a classification head with features from the fine-tuned Sentence Transformer.
Model Details
Model Description
- Model Type: SetFit
- Sentence Transformer body: projecte-aina/ST-NLI-ca_paraphrase-multilingual-mpnet-base
- Classification head: a LogisticRegression instance
- Maximum Sequence Length: 128 tokens
- Number of Classes: 20 classes
Model Sources
- Repository: SetFit on GitHub
- Paper: Efficient Few-Shot Learning Without Prompts
- Blogpost: SetFit: Efficient Few-Shot Learning Without Prompts
Model Labels
Label | Examples |
---|---|
0 |
|
1 |
|
2 |
|
3 |
|
4 |
|
5 |
|
6 |
|
7 |
|
8 |
|
9 |
|
10 |
|
11 |
|
12 |
|
13 |
|
14 |
|
15 |
|
16 |
|
17 |
|
18 |
|
19 |
|
Uses
Direct Use for Inference
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("adriansanz/fs_setfit_dummy")
# Run inference
preds = model("Suggeriria que es realitzessin campanyes de recompensa per incentivar els ciutadans a informar de fuites d'aigua, oferint descomptes en la factura d'aigua o altres incentius.")
Training Details
Training Set Metrics
Training set | Min | Median | Max |
---|---|---|---|
Word count | 4 | 4.85 | 8 |
Label | Training Sample Count |
---|---|
0 | 8 |
1 | 8 |
2 | 8 |
3 | 8 |
4 | 8 |
5 | 8 |
6 | 8 |
7 | 8 |
8 | 8 |
9 | 8 |
10 | 8 |
11 | 8 |
12 | 8 |
13 | 8 |
14 | 8 |
15 | 8 |
16 | 8 |
17 | 8 |
18 | 8 |
19 | 8 |
Training Hyperparameters
- batch_size: (16, 16)
- num_epochs: (3, 3)
- max_steps: -1
- sampling_strategy: oversampling
- body_learning_rate: (2e-05, 1e-05)
- head_learning_rate: 0.01
- loss: CosineSimilarityLoss
- distance_metric: cosine_distance
- margin: 0.25
- end_to_end: False
- use_amp: False
- warmup_proportion: 0.1
- seed: 42
- eval_max_steps: -1
- load_best_model_at_end: True
Training Results
Epoch | Step | Training Loss | Validation Loss |
---|---|---|---|
0.0007 | 1 | 0.1362 | - |
0.0329 | 50 | 0.0344 | - |
0.0658 | 100 | 0.0017 | - |
0.0987 | 150 | 0.0013 | - |
0.1316 | 200 | 0.0013 | - |
0.1645 | 250 | 0.0007 | - |
0.1974 | 300 | 0.0004 | - |
0.2303 | 350 | 0.0004 | - |
0.2632 | 400 | 0.0006 | - |
0.2961 | 450 | 0.0005 | - |
0.3289 | 500 | 0.0003 | - |
0.3618 | 550 | 0.0005 | - |
0.3947 | 600 | 0.0006 | - |
0.4276 | 650 | 0.0004 | - |
0.4605 | 700 | 0.0003 | - |
0.4934 | 750 | 0.0001 | - |
0.5263 | 800 | 0.0002 | - |
0.5592 | 850 | 0.0002 | - |
0.5921 | 900 | 0.0002 | - |
0.625 | 950 | 0.0002 | - |
0.6579 | 1000 | 0.0002 | - |
0.6908 | 1050 | 0.0002 | - |
0.7237 | 1100 | 0.0002 | - |
0.7566 | 1150 | 0.0002 | - |
0.7895 | 1200 | 0.0002 | - |
0.8224 | 1250 | 0.0003 | - |
0.8553 | 1300 | 0.0002 | - |
0.8882 | 1350 | 0.0001 | - |
0.9211 | 1400 | 0.0001 | - |
0.9539 | 1450 | 0.0002 | - |
0.9868 | 1500 | 0.0002 | - |
1.0 | 1520 | - | 0.1669 |
1.0197 | 1550 | 0.0002 | - |
1.0526 | 1600 | 0.0001 | - |
1.0855 | 1650 | 0.0003 | - |
1.1184 | 1700 | 0.0002 | - |
1.1513 | 1750 | 0.0002 | - |
1.1842 | 1800 | 0.0001 | - |
1.2171 | 1850 | 0.0002 | - |
1.25 | 1900 | 0.0003 | - |
1.2829 | 1950 | 0.0002 | - |
1.3158 | 2000 | 0.0001 | - |
1.3487 | 2050 | 0.0002 | - |
1.3816 | 2100 | 0.0001 | - |
1.4145 | 2150 | 0.0001 | - |
1.4474 | 2200 | 0.0001 | - |
1.4803 | 2250 | 0.0002 | - |
1.5132 | 2300 | 0.0002 | - |
1.5461 | 2350 | 0.0002 | - |
1.5789 | 2400 | 0.0001 | - |
1.6118 | 2450 | 0.0001 | - |
1.6447 | 2500 | 0.0002 | - |
1.6776 | 2550 | 0.0002 | - |
1.7105 | 2600 | 0.0002 | - |
1.7434 | 2650 | 0.0001 | - |
1.7763 | 2700 | 0.0001 | - |
1.8092 | 2750 | 0.0001 | - |
1.8421 | 2800 | 0.0001 | - |
1.875 | 2850 | 0.0001 | - |
1.9079 | 2900 | 0.0001 | - |
1.9408 | 2950 | 0.0001 | - |
1.9737 | 3000 | 0.0001 | - |
2.0 | 3040 | - | 0.1629 |
2.0066 | 3050 | 0.0001 | - |
2.0395 | 3100 | 0.0001 | - |
2.0724 | 3150 | 0.0001 | - |
2.1053 | 3200 | 0.0001 | - |
2.1382 | 3250 | 0.0001 | - |
2.1711 | 3300 | 0.0001 | - |
2.2039 | 3350 | 0.0001 | - |
2.2368 | 3400 | 0.0001 | - |
2.2697 | 3450 | 0.0001 | - |
2.3026 | 3500 | 0.0002 | - |
2.3355 | 3550 | 0.0001 | - |
2.3684 | 3600 | 0.0001 | - |
2.4013 | 3650 | 0.0001 | - |
2.4342 | 3700 | 0.0001 | - |
2.4671 | 3750 | 0.0001 | - |
2.5 | 3800 | 0.0001 | - |
2.5329 | 3850 | 0.0001 | - |
2.5658 | 3900 | 0.0001 | - |
2.5987 | 3950 | 0.0 | - |
2.6316 | 4000 | 0.0 | - |
2.6645 | 4050 | 0.0001 | - |
2.6974 | 4100 | 0.0 | - |
2.7303 | 4150 | 0.0001 | - |
2.7632 | 4200 | 0.0001 | - |
2.7961 | 4250 | 0.0001 | - |
2.8289 | 4300 | 0.0001 | - |
2.8618 | 4350 | 0.0001 | - |
2.8947 | 4400 | 0.0001 | - |
2.9276 | 4450 | 0.0001 | - |
2.9605 | 4500 | 0.0001 | - |
2.9934 | 4550 | 0.0 | - |
3.0 | 4560 | - | 0.1625 |
- The bold row denotes the saved checkpoint.
Framework Versions
- Python: 3.10.12
- SetFit: 1.0.3
- Sentence Transformers: 3.0.0
- Transformers: 4.39.0
- PyTorch: 2.3.0+cu121
- Datasets: 2.19.1
- Tokenizers: 0.15.2
Citation
BibTeX
@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}
}
- Downloads last month
- 19
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.