SetFit with Alibaba-NLP/gte-multilingual-base

This is a SetFit model trained on the tmp-org/edeka-dataset-ctx-1 dataset that can be used for Text Classification. This SetFit model uses Alibaba-NLP/gte-multilingual-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: Alibaba-NLP/gte-multilingual-base
Classification head: a LogisticRegression instance
Maximum Sequence Length: 8192 tokens
Number of Classes: 24 classes
Training Dataset: tmp-org/edeka-dataset-ctx-1

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
Start_Start	'Sparen ist App Sache () [TextView
Kasse_Aktivierte Coupons	' (Aktivierte Coupons) [View
Other_Prospekt	'Wir freuen uns über deine Bewertung oder dein Feedback zur EDEKA App. () [TextView
Other_Unknown	'Prospekt () [TextView
Sparen_Coupons	'Coupons () [TextView
Sparen_Angebote	'Coupons () [TextView
Kasse_Kasse	'Coupons () [TextView
Prämien_Prämien	'Coupons () [TextView
Other_Treueaktionen	'3.79 () [TextView
Other_Neuigkeiten	'Im ausgewählten Markt verfügbar\nEDEKA Schöck () [TextView
Other_Produktherkunft	'Pura Raza Española Stute () [TextView
Einkaufsliste_Einkaufsliste	'sortieren nach () [TextView
Other_Loading	'2 / 12 () [TextView
Other_Menu	'sortieren nach () [TextView
Other_Kassenbons	' (Tabelle schließen) [View
Kasse_Unknown	' (Tabelle schließen) [View
Other_Coupon details	'Coca-Cola 0,33l DPG\nFanta Exotic 0,33l DPG\nCoca-Cola Zero Sugar 0,33l DPG\nFanta Orange 0,33l DPG\nSprite Zitrone-Limette 0,33l DPG\nMezzo Mix 0,33l DPG\nCoca-Cola Cherry 0,33l DPG\nFanta Mango Dragonfruit 0,33l DPG\nSprite Zitrone-Limette Zero 0,33l DPG\nCoca-Cola Light 0,33l DPG\nFanta Lemon Elderflower 0,33l DPG\nFanta Zero Sugar Orange 0,33l DPG\nFanta Lemon 0,33l DPG\nCoca-Cola Lemon 0,33l DPG\nCoca-Cola Zero Koffeinfrei 0,33l DPG\nMezzo Mix Zero 0,33l DPG\nMezzo Mix Berry Love 0,33l DPG\nCoca-Cola Vanilla 0,33l DPG\nFanta Cassis 0,33l DPG\nCoca-Cola Zero Sugar 0,33l DPG\nFanta Zero Sugar Tutti Frutti 0,33l DPG\nFanta Zero Forest Berries 0,33l DPG\n\n5000112602029\n5000112514124\n5000112604450\n5000112604726\n5000112604740\n5000112604696\n5000112604467\n5000112515848\n5000112636420\n5000112604627\n5000112514100\n5000112668766\n5000112510379\n5000112684667\n5000112668872\n5000112672077\n5000112677430\n5000112682205\n5000112680546\n5449000275165\n5000112688801\n5000112686869 () [TextView
Kasse_Mobil bezahlen	' (Tabelle schließen) [View
Start_Loading	'Gratisartikel im November: Gut&Günstig – Classic Spülmittel Apfelduft () [TextView
Sparen_Loading	'Zukünftige Treueaktionen () [TextView
Other_Marktsuche	'Aktuelle Treueaktionen () [TextView
Other_Other	'Obst & Gemüse () [TextView
Other_Code einlösen	'Weitere Funktionen () [TextView
Kasse_Loading	'Mobil bezahlen per App () [TextView

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("tmp-org/tmp_cv_model_2025_09_08_0")
# Run inference
preds = model("Prospekt () [TextView|MainActivity] | 1 / 12 () [TextView|MainActivity]

[SELECTED START]

2 / 12 () [TextView|MainActivity]
[CONTEXT SEPARATOR]
2 / 12 () [TextView|MainActivity]")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	9	270.1332	711

Label	Training Sample Count
Einkaufsliste_Einkaufsliste	31
Kasse_Aktivierte Coupons	40
Kasse_Kasse	17
Kasse_Loading	2
Kasse_Mobil bezahlen	5
Kasse_Unknown	1
Other_Code einlösen	1
Other_Coupon details	32
Other_Kassenbons	8
Other_Loading	1
Other_Marktsuche	2
Other_Menu	40
Other_Neuigkeiten	6
Other_Other	1
Other_Produktherkunft	11
Other_Prospekt	33
Other_Treueaktionen	36
Other_Unknown	10
Prämien_Prämien	38
Sparen_Angebote	40
Sparen_Coupons	40
Sparen_Loading	3
Start_Loading	5
Start_Start	40

Training Hyperparameters

batch_size: (4, 4)
num_epochs: (1, 1)
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: 0.1
l2_weight: 0.01
seed: 4242
eval_max_steps: -1
load_best_model_at_end: False

Training Results

Epoch	Step	Training Loss	Validation Loss
0.0003	1	0.1732	-
0.0134	50	0.2443	-
0.0268	100	0.1933	-
0.0402	150	0.1816	-
0.0535	200	0.1304	-
0.0669	250	0.1268	-
0.0803	300	0.122	-
0.0937	350	0.1228	-
0.1071	400	0.1263	-
0.1205	450	0.0822	-
0.1339	500	0.0648	-
0.1473	550	0.1123	-
0.1606	600	0.08	-
0.1740	650	0.0927	-
0.1874	700	0.0722	-
0.2008	750	0.0662	-
0.2142	800	0.098	-
0.2276	850	0.0456	-
0.2410	900	0.0407	-
0.2544	950	0.0673	-
0.2677	1000	0.0408	-
0.2811	1050	0.0634	-
0.2945	1100	0.0608	-
0.3079	1150	0.0745	-
0.3213	1200	0.0379	-
0.3347	1250	0.0267	-
0.3481	1300	0.0393	-
0.3614	1350	0.0342	-
0.3748	1400	0.0539	-
0.3882	1450	0.0343	-
0.4016	1500	0.0331	-
0.4150	1550	0.0182	-
0.4284	1600	0.0486	-
0.4418	1650	0.0395	-
0.4552	1700	0.0462	-
0.4685	1750	0.0264	-
0.4819	1800	0.0341	-
0.4953	1850	0.0296	-
0.5087	1900	0.0262	-
0.5221	1950	0.0527	-
0.5355	2000	0.0446	-
0.5489	2050	0.0311	-
0.5622	2100	0.025	-
0.5756	2150	0.0251	-
0.5890	2200	0.0224	-
0.6024	2250	0.0469	-
0.6158	2300	0.0336	-
0.6292	2350	0.0258	-
0.6426	2400	0.0326	-
0.6560	2450	0.027	-
0.6693	2500	0.036	-
0.6827	2550	0.0286	-
0.6961	2600	0.0273	-
0.7095	2650	0.0288	-
0.7229	2700	0.0267	-
0.7363	2750	0.0412	-
0.7497	2800	0.0202	-
0.7631	2850	0.0244	-
0.7764	2900	0.0359	-
0.7898	2950	0.0377	-
0.8032	3000	0.0302	-
0.8166	3050	0.0192	-
0.8300	3100	0.0296	-
0.8434	3150	0.0292	-
0.8568	3200	0.0317	-
0.8701	3250	0.0246	-
0.8835	3300	0.0096	-
0.8969	3350	0.0306	-
0.9103	3400	0.0198	-
0.9237	3450	0.0188	-
0.9371	3500	0.0253	-
0.9505	3550	0.0292	-
0.9639	3600	0.04	-
0.9772	3650	0.0298	-
0.9906	3700	0.0184	-

Framework Versions

Python: 3.12.6
SetFit: 1.1.2
Sentence Transformers: 5.2.2
Transformers: 4.57.1
PyTorch: 2.10.0+cu128
Datasets: 3.6.0
Tokenizers: 0.22.2

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: 31

Safetensors

Model size

0.3B params

Tensor type

F32

Model tree for tmp-org/tmp_cv_model_2025_09_08_0

Base model

Alibaba-NLP/gte-multilingual-base

Finetuned

(98)

this model

Paper for tmp-org/tmp_cv_model_2025_09_08_0

Efficient Few-Shot Learning Without Prompts

Paper • 2209.11055 • Published Sep 22, 2022 • 4