Input Classifier

This is a SetFit model that can be used for Text Classification. This SetFit model uses jaimevera1107/all-MiniLM-L6-v2-similarity-es 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: jaimevera1107/all-MiniLM-L6-v2-similarity-es
Classification head: a LogisticRegression instance
Maximum Sequence Length: 256 tokens
Number of Classes: 4 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
request	'lencería necesaria' 'material necesario para hoy' 'terminé la habitación 234'
conversation	'buena noche' 'Qué pasa, tío, ¿todo bien?' 'Buenas, ¿cómo va la cosa?!'
help	'ayuda por favor' 'Ayuda que no sé que puedo hacer' 'Hola, que puedo hacer'
censorship	'Eres un completo inútil, no sirves para nada' 'Siempre diciendo estupideces, mejor cállate' 'Tu sola existencia es una vergüenza'

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("monentiadev/es-input-classifier")
# Run inference
preds = model("Hola")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	1	5.0723	38

Label	Training Sample Count
censorship	407
conversation	137
help	274
request	552

Training Hyperparameters

batch_size: (128, 128)
num_epochs: (3, 3)
max_steps: -1
sampling_strategy: oversampling
num_iterations: 20
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
0.0023	1	0.3161	-
0.1166	50	0.2857	-
0.2331	100	0.2158	-
0.3497	150	0.1581	-
0.4662	200	0.0878	-
0.5828	250	0.0299	-
0.6993	300	0.0124	-
0.8159	350	0.0083	-
0.9324	400	0.006	-
1.0490	450	0.0038	-
1.1655	500	0.0027	-
1.2821	550	0.0027	-
1.3986	600	0.0017	-
1.5152	650	0.0016	-
1.6317	700	0.0013	-
1.7483	750	0.0012	-
1.8648	800	0.0012	-
1.9814	850	0.001	-
2.0979	900	0.001	-
2.2145	950	0.0011	-
2.3310	1000	0.0009	-
2.4476	1050	0.0008	-
2.5641	1100	0.0009	-
2.6807	1150	0.0008	-
2.7972	1200	0.0008	-
2.9138	1250	0.0007	-

Framework Versions

Python: 3.10.0
SetFit: 1.1.2
Sentence Transformers: 5.0.0
Transformers: 4.53.1
PyTorch: 2.7.1+cu126
Datasets: 2.19.2
Tokenizers: 0.21.2

monentiadev
/

es-input-classifier