lucasflins/first_model_setfit

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

This is a SetFit model that can be used for Text Classification. This SetFit model uses sentence-transformers/paraphrase-mpnet-base-v2 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:

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/paraphrase-mpnet-base-v2
• Classification head: a LogisticRegression instance
• Maximum Sequence Length: 512 tokens
• Number of Classes: 2 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
positive	'tintas esmalte a oleo' 'jogo xicaras' 'cerveja'
negative	'receptor gbr' 'foco roubado' 'unitv v10'

Evaluation

Metrics

Label	Accuracy
all	0.8644

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("lucasflins/first_model_setfit")
# Run inference
preds = model("sulgador d clitoris")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	1	2.9880	20

Label	Training Sample Count
negative	458
positive	545

Training Hyperparameters

• batch_size: (512, 512)
• num_epochs: (5, 5)
• 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
• l2_weight: 0.01
• seed: 42
• eval_max_steps: -1
• load_best_model_at_end: True

Training Results

Epoch	Step	Training Loss	Validation Loss
0.0010	1	0.3235	-
0.0504	50	0.2952	-
0.1008	100	0.2548	-
0.1512	150	0.2506	-
0.2016	200	0.247	-
0.2520	250	0.2168	-
0.3024	300	0.081	-
0.3528	350	0.0229	-
0.4032	400	0.0114	-
0.4536	450	0.0086	-
0.5040	500	0.0091	-
0.5544	550	0.0059	-
0.6048	600	0.0041	-
0.6552	650	0.0027	-
0.7056	700	0.0016	-
0.7560	750	0.0011	-
0.8065	800	0.001	-
0.8569	850	0.0011	-
0.9073	900	0.0014	-
0.9577	950	0.0009	-
1.0	992	-	0.2117
1.0081	1000	0.0007	-
1.0585	1050	0.0008	-
1.1089	1100	0.0005	-
1.1593	1150	0.0006	-
1.2097	1200	0.0007	-
1.2601	1250	0.0005	-
1.3105	1300	0.0004	-
1.3609	1350	0.0007	-
1.4113	1400	0.0007	-
1.4617	1450	0.0004	-
1.5121	1500	0.0005	-
1.5625	1550	0.0005	-
1.6129	1600	0.0005	-
1.6633	1650	0.0004	-
1.7137	1700	0.0003	-
1.7641	1750	0.0005	-
1.8145	1800	0.0006	-
1.8649	1850	0.0002	-
1.9153	1900	0.0003	-
1.9657	1950	0.0003	-
2.0	1984	-	0.2333
2.0161	2000	0.0003	-
2.0665	2050	0.0005	-
2.1169	2100	0.0005	-
2.1673	2150	0.0003	-
2.2177	2200	0.0002	-
2.2681	2250	0.0004	-
2.3185	2300	0.0006	-
2.3690	2350	0.0005	-
2.4194	2400	0.0001	-
2.4698	2450	0.0005	-
2.5202	2500	0.0006	-
2.5706	2550	0.0004	-
2.6210	2600	0.0004	-
2.6714	2650	0.0004	-
2.7218	2700	0.0001	-
2.7722	2750	0.0001	-
2.8226	2800	0.0001	-
2.8730	2850	0.0001	-
2.9234	2900	0.0001	-
2.9738	2950	0.0002	-
3.0	2976	-	0.2406
3.0242	3000	0.0001	-
3.0746	3050	0.0001	-
3.125	3100	0.0002	-
3.1754	3150	0.0002	-
3.2258	3200	0.0001	-
3.2762	3250	0.0001	-
3.3266	3300	0.0002	-
3.3770	3350	0.0001	-
3.4274	3400	0.0003	-
3.4778	3450	0.0002	-
3.5282	3500	0.0001	-
3.5786	3550	0.0001	-
3.6290	3600	0.0001	-
3.6794	3650	0.0001	-
3.7298	3700	0.0001	-
3.7802	3750	0.0001	-
3.8306	3800	0.0002	-
3.8810	3850	0.0003	-
3.9315	3900	0.0001	-
3.9819	3950	0.0001	-
4.0	3968	-	0.2248
4.0323	4000	0.0002	-
4.0827	4050	0.0001	-
4.1331	4100	0.0001	-
4.1835	4150	0.0002	-
4.2339	4200	0.0001	-
4.2843	4250	0.0001	-
4.3347	4300	0.0001	-
4.3851	4350	0.0001	-
4.4355	4400	0.0001	-
4.4859	4450	0.0001	-
4.5363	4500	0.0001	-
4.5867	4550	0.0001	-
4.6371	4600	0.0001	-
4.6875	4650	0.0001	-
4.7379	4700	0.0001	-
4.7883	4750	0.0001	-
4.8387	4800	0.0001	-
4.8891	4850	0.0001	-
4.9395	4900	0.0001	-
4.9899	4950	0.0001	-
5.0	4960	-	0.2303

Framework Versions

• Python: 3.12.11
• SetFit: 1.1.3
• Sentence Transformers: 5.1.2
• Transformers: 4.56.2
• PyTorch: 2.7.1+cu128
• Datasets: 4.1.1
• Tokenizers: 0.22.1

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}
}