GIZ/VULNERABILITY-multilabel-mpnet

SetFit with sentence-transformers/all-mpnet-base-v2

This is a SetFit model that can be used for Text Classification. This SetFit model uses sentence-transformers/all-mpnet-base-v2 as the Sentence Transformer embedding model. A SetFitHead instance is used for classification.

The model has been trained using an efficient few-shot learning technique that involves:

1. Fine-tuning a Sentence Transformer with contrastive learning.
2. Training a classification head with features from the fine-tuned Sentence Transformer.

Model Details

Model Description

• Model Type: SetFit
• Sentence Transformer body: sentence-transformers/all-mpnet-base-v2
• Classification head: a SetFitHead instance
• Maximum Sequence Length: 384 tokens
• Number of Classes: 18 classes

Model Sources

• Repository: SetFit on GitHub
• Paper: Efficient Few-Shot Learning Without Prompts
• Blogpost: SetFit: Efficient Few-Shot Learning Without Prompts

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("leavoigt/vulnerability_multilabel_updated")
# Run inference
preds = model("Workers in the formal sector. Formal sector workers also face economic risks. A number of them experience income instability due to contractualization, retrenchment, and firm closures. In 2014, contractual workers accounted for 22 percent of the total 4.5 million workers employed in establishments with 20 or more employees.")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	21	72.6472	238

Training Hyperparameters

• batch_size: (16, 2)
• num_epochs: (1, 0)
• max_steps: -1
• sampling_strategy: undersampling
• 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: 0.01
• seed: 42
• eval_max_steps: -1
• load_best_model_at_end: False

Training Results

Epoch	Step	Training Loss	Validation Loss
0.0006	1	0.1906	-
0.0316	50	0.1275	0.1394
0.0631	100	0.0851	0.1247
0.0947	150	0.0959	0.1269
0.1263	200	0.1109	0.1179
0.1578	250	0.0923	0.1354
0.1894	300	0.063	0.1292
0.2210	350	0.0555	0.1326
0.2525	400	0.0362	0.1127
0.2841	450	0.0582	0.132
0.3157	500	0.0952	0.1339
0.3472	550	0.0793	0.1171
0.3788	600	0.059	0.1187
0.4104	650	0.0373	0.1131
0.4419	700	0.0593	0.1144
0.4735	750	0.0405	0.1174
0.5051	800	0.0284	0.1196
0.5366	850	0.0329	0.1116
0.5682	900	0.0895	0.1193
0.5997	950	0.0576	0.1159
0.6313	1000	0.0385	0.1203
0.6629	1050	0.0842	0.1195
0.6944	1100	0.0274	0.113
0.7260	1150	0.0226	0.1137
0.7576	1200	0.0276	0.1204
0.7891	1250	0.0355	0.1163
0.8207	1300	0.077	0.1161
0.8523	1350	0.0735	0.1135
0.8838	1400	0.0357	0.1175
0.9154	1450	0.0313	0.1207
0.9470	1500	0.0241	0.1159
0.9785	1550	0.0339	0.1161

Framework Versions

• Python: 3.10.12
• SetFit: 1.0.3
• Sentence Transformers: 2.3.1
• Transformers: 4.38.1
• PyTorch: 2.1.0+cu121
• Datasets: 2.3.0
• 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}
}