SetFit with intfloat/multilingual-e5-base
This is a SetFit model that can be used for Text Classification. This SetFit model uses intfloat/multilingual-e5-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: intfloat/multilingual-e5-base
- Classification head: a LogisticRegression instance
- Maximum Sequence Length: 512 tokens
- Number of Classes: 12 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 |
---|---|
Business |
|
Politics |
|
Entertainment |
|
Crime |
|
Science |
|
Lifestyle |
|
Education |
|
Sports |
|
Health |
|
General News |
|
Technology |
|
Religion |
|
Evaluation
Metrics
Label | Accuracy |
---|---|
all | 0.8222 |
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("setfit_model_id")
# Run inference
preds = model("Bavarian PM calls to stop refugee payments for Ukrainians Ambassador responds. Read more")
Training Details
Training Set Metrics
Training set | Min | Median | Max |
---|---|---|---|
Word count | 2 | 42.6906 | 454 |
Label | Training Sample Count |
---|---|
Business | 346 |
Sports | 244 |
Politics | 210 |
Lifestyle | 186 |
General News | 186 |
Entertainment | 150 |
Crime | 98 |
Technology | 71 |
Health | 70 |
Science | 30 |
Religion | 13 |
Education | 12 |
Training Hyperparameters
- batch_size: (16, 16)
- num_epochs: (5, 5)
- max_steps: -1
- sampling_strategy: oversampling
- num_iterations: 10
- 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
- l2_weight: 0.01
- seed: 42
- eval_max_steps: -1
- load_best_model_at_end: False
Training Results
Epoch | Step | Training Loss | Validation Loss |
---|---|---|---|
0.0005 | 1 | 0.1861 | - |
0.0248 | 50 | 0.3204 | - |
0.0495 | 100 | 0.2765 | - |
0.0743 | 150 | 0.2462 | - |
0.0990 | 200 | 0.2192 | - |
0.1238 | 250 | 0.1755 | - |
0.1485 | 300 | 0.135 | - |
0.1733 | 350 | 0.1135 | - |
0.1980 | 400 | 0.092 | - |
0.2228 | 450 | 0.0885 | - |
0.2475 | 500 | 0.0739 | - |
0.2723 | 550 | 0.0762 | - |
0.2970 | 600 | 0.0688 | - |
0.3218 | 650 | 0.0633 | - |
0.3465 | 700 | 0.0535 | - |
0.3713 | 750 | 0.0363 | - |
0.3960 | 800 | 0.0388 | - |
0.4208 | 850 | 0.0339 | - |
0.4455 | 900 | 0.0265 | - |
0.4703 | 950 | 0.0344 | - |
0.4950 | 1000 | 0.016 | - |
0.5198 | 1050 | 0.0231 | - |
0.5446 | 1100 | 0.0152 | - |
0.5693 | 1150 | 0.0118 | - |
0.5941 | 1200 | 0.0102 | - |
0.6188 | 1250 | 0.0089 | - |
0.6436 | 1300 | 0.0125 | - |
0.6683 | 1350 | 0.0082 | - |
0.6931 | 1400 | 0.004 | - |
0.7178 | 1450 | 0.004 | - |
0.7426 | 1500 | 0.0062 | - |
0.7673 | 1550 | 0.004 | - |
0.7921 | 1600 | 0.0039 | - |
0.8168 | 1650 | 0.0111 | - |
0.8416 | 1700 | 0.0024 | - |
0.8663 | 1750 | 0.0047 | - |
0.8911 | 1800 | 0.0013 | - |
0.9158 | 1850 | 0.0023 | - |
0.9406 | 1900 | 0.0039 | - |
0.9653 | 1950 | 0.0036 | - |
0.9901 | 2000 | 0.004 | - |
1.0149 | 2050 | 0.0007 | - |
1.0396 | 2100 | 0.001 | - |
1.0644 | 2150 | 0.0029 | - |
1.0891 | 2200 | 0.0005 | - |
1.1139 | 2250 | 0.0005 | - |
1.1386 | 2300 | 0.0006 | - |
1.1634 | 2350 | 0.0003 | - |
1.1881 | 2400 | 0.0002 | - |
1.2129 | 2450 | 0.0018 | - |
1.2376 | 2500 | 0.0013 | - |
1.2624 | 2550 | 0.0039 | - |
1.2871 | 2600 | 0.0025 | - |
1.3119 | 2650 | 0.0025 | - |
1.3366 | 2700 | 0.0013 | - |
1.3614 | 2750 | 0.0017 | - |
1.3861 | 2800 | 0.0005 | - |
1.4109 | 2850 | 0.0012 | - |
1.4356 | 2900 | 0.0002 | - |
1.4604 | 2950 | 0.0006 | - |
1.4851 | 3000 | 0.0017 | - |
1.5099 | 3050 | 0.0004 | - |
1.5347 | 3100 | 0.0002 | - |
1.5594 | 3150 | 0.0015 | - |
1.5842 | 3200 | 0.0002 | - |
1.6089 | 3250 | 0.0002 | - |
1.6337 | 3300 | 0.0023 | - |
1.6584 | 3350 | 0.0025 | - |
1.6832 | 3400 | 0.0002 | - |
1.7079 | 3450 | 0.0006 | - |
1.7327 | 3500 | 0.0006 | - |
1.7574 | 3550 | 0.0014 | - |
1.7822 | 3600 | 0.0003 | - |
1.8069 | 3650 | 0.0024 | - |
1.8317 | 3700 | 0.0003 | - |
1.8564 | 3750 | 0.001 | - |
1.8812 | 3800 | 0.0005 | - |
1.9059 | 3850 | 0.0014 | - |
1.9307 | 3900 | 0.0007 | - |
1.9554 | 3950 | 0.0016 | - |
1.9802 | 4000 | 0.0013 | - |
2.0050 | 4050 | 0.0007 | - |
2.0297 | 4100 | 0.001 | - |
2.0545 | 4150 | 0.0005 | - |
2.0792 | 4200 | 0.0002 | - |
2.1040 | 4250 | 0.0001 | - |
2.1287 | 4300 | 0.0003 | - |
2.1535 | 4350 | 0.0001 | - |
2.1782 | 4400 | 0.0009 | - |
2.2030 | 4450 | 0.0002 | - |
2.2277 | 4500 | 0.0004 | - |
2.2525 | 4550 | 0.0003 | - |
2.2772 | 4600 | 0.0001 | - |
2.3020 | 4650 | 0.0001 | - |
2.3267 | 4700 | 0.0011 | - |
2.3515 | 4750 | 0.0016 | - |
2.3762 | 4800 | 0.0004 | - |
2.4010 | 4850 | 0.0002 | - |
2.4257 | 4900 | 0.0001 | - |
2.4505 | 4950 | 0.0004 | - |
2.4752 | 5000 | 0.0001 | - |
2.5 | 5050 | 0.0002 | - |
2.5248 | 5100 | 0.0017 | - |
2.5495 | 5150 | 0.0002 | - |
2.5743 | 5200 | 0.0001 | - |
2.5990 | 5250 | 0.0013 | - |
2.6238 | 5300 | 0.0014 | - |
2.6485 | 5350 | 0.0001 | - |
2.6733 | 5400 | 0.0001 | - |
2.6980 | 5450 | 0.0001 | - |
2.7228 | 5500 | 0.0001 | - |
2.7475 | 5550 | 0.0001 | - |
2.7723 | 5600 | 0.0001 | - |
2.7970 | 5650 | 0.0001 | - |
2.8218 | 5700 | 0.0 | - |
2.8465 | 5750 | 0.0 | - |
2.8713 | 5800 | 0.0012 | - |
2.8960 | 5850 | 0.0001 | - |
2.9208 | 5900 | 0.0001 | - |
2.9455 | 5950 | 0.0003 | - |
2.9703 | 6000 | 0.0001 | - |
2.9950 | 6050 | 0.0001 | - |
3.0198 | 6100 | 0.0 | - |
3.0446 | 6150 | 0.0 | - |
3.0693 | 6200 | 0.0 | - |
3.0941 | 6250 | 0.0 | - |
3.1188 | 6300 | 0.0 | - |
3.1436 | 6350 | 0.0 | - |
3.1683 | 6400 | 0.0 | - |
3.1931 | 6450 | 0.0001 | - |
3.2178 | 6500 | 0.0001 | - |
3.2426 | 6550 | 0.0001 | - |
3.2673 | 6600 | 0.0 | - |
3.2921 | 6650 | 0.0 | - |
3.3168 | 6700 | 0.0 | - |
3.3416 | 6750 | 0.0 | - |
3.3663 | 6800 | 0.0 | - |
3.3911 | 6850 | 0.0012 | - |
3.4158 | 6900 | 0.0013 | - |
3.4406 | 6950 | 0.0001 | - |
3.4653 | 7000 | 0.001 | - |
3.4901 | 7050 | 0.0001 | - |
3.5149 | 7100 | 0.0002 | - |
3.5396 | 7150 | 0.0002 | - |
3.5644 | 7200 | 0.0001 | - |
3.5891 | 7250 | 0.0001 | - |
3.6139 | 7300 | 0.0002 | - |
3.6386 | 7350 | 0.0001 | - |
3.6634 | 7400 | 0.0001 | - |
3.6881 | 7450 | 0.0013 | - |
3.7129 | 7500 | 0.0001 | - |
3.7376 | 7550 | 0.0 | - |
3.7624 | 7600 | 0.0 | - |
3.7871 | 7650 | 0.0 | - |
3.8119 | 7700 | 0.0 | - |
3.8366 | 7750 | 0.0 | - |
3.8614 | 7800 | 0.0 | - |
3.8861 | 7850 | 0.0 | - |
3.9109 | 7900 | 0.0 | - |
3.9356 | 7950 | 0.0 | - |
3.9604 | 8000 | 0.0 | - |
3.9851 | 8050 | 0.0 | - |
4.0099 | 8100 | 0.0001 | - |
4.0347 | 8150 | 0.0 | - |
4.0594 | 8200 | 0.0 | - |
4.0842 | 8250 | 0.0 | - |
4.1089 | 8300 | 0.0 | - |
4.1337 | 8350 | 0.0 | - |
4.1584 | 8400 | 0.0 | - |
4.1832 | 8450 | 0.0 | - |
4.2079 | 8500 | 0.0 | - |
4.2327 | 8550 | 0.0 | - |
4.2574 | 8600 | 0.0 | - |
4.2822 | 8650 | 0.0 | - |
4.3069 | 8700 | 0.0 | - |
4.3317 | 8750 | 0.0 | - |
4.3564 | 8800 | 0.0 | - |
4.3812 | 8850 | 0.0 | - |
4.4059 | 8900 | 0.0 | - |
4.4307 | 8950 | 0.0 | - |
4.4554 | 9000 | 0.0 | - |
4.4802 | 9050 | 0.0 | - |
4.5050 | 9100 | 0.0 | - |
4.5297 | 9150 | 0.0 | - |
4.5545 | 9200 | 0.0 | - |
4.5792 | 9250 | 0.0008 | - |
4.6040 | 9300 | 0.0001 | - |
4.6287 | 9350 | 0.0 | - |
4.6535 | 9400 | 0.0 | - |
4.6782 | 9450 | 0.0 | - |
4.7030 | 9500 | 0.0 | - |
4.7277 | 9550 | 0.0013 | - |
4.7525 | 9600 | 0.0001 | - |
4.7772 | 9650 | 0.0 | - |
4.8020 | 9700 | 0.0001 | - |
4.8267 | 9750 | 0.0 | - |
4.8515 | 9800 | 0.0 | - |
4.8762 | 9850 | 0.0 | - |
4.9010 | 9900 | 0.0 | - |
4.9257 | 9950 | 0.0001 | - |
4.9505 | 10000 | 0.0 | - |
4.9752 | 10050 | 0.0 | - |
5.0 | 10100 | 0.0001 | - |
Framework Versions
- Python: 3.10.12
- SetFit: 1.1.3
- Sentence Transformers: 5.1.0
- Transformers: 4.55.2
- PyTorch: 2.7.1+cu118
- Datasets: 4.0.0
- Tokenizers: 0.21.4
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
- 66
Model tree for alunadiderot/setfit-e5-base-category-classifier_v2
Base model
intfloat/multilingual-e5-base