metadata

base_model: BAAI/bge-small-en-v1.5
language: en
library_name: setfit
license: apache-2.0
metrics:
  - '0'
  - '1'
  - accuracy
  - macro avg
  - weighted avg
pipeline_tag: text-classification
tags:
  - setfit
  - sentence-transformers
  - text-classification
  - generated_from_setfit_trainer
widget:
  - text: >-
      the nerves,blood vessels, and glands are located in which layer of the
      skin
  - text: Where would you put refuse if you do not want it to exist any more?
  - text: Obesity can cause resistance to which hormone?
  - text: Referees
  - text: where does the water at niagra falls come from
inference: true
model-index:
  - name: SetFit with BAAI/bge-small-en-v1.5 on Health Information Needs
    results:
      - task:
          type: text-classification
          name: Text Classification
        dataset:
          name: Health Information Needs
          type: unknown
          split: test
        metrics:
          - type: '0'
            value:
              precision: 0.36527485731450887
              recall: 0.990228013029316
              f1-score: 0.5336844415185429
              support: 1228
            name: '0'
          - type: '1'
            value:
              precision: 0.994328922495274
              recall: 0.4989328906805786
              f1-score: 0.6644560240012632
              support: 4217
            name: '1'
          - type: accuracy
            value: 0.6097337006427915
            name: Accuracy
          - type: macro avg
            value:
              precision: 0.6798018899048914
              recall: 0.7445804518549473
              f1-score: 0.5990702327599031
              support: 5445
            name: Macro Avg
          - type: weighted avg
            value:
              precision: 0.8524596126620364
              recall: 0.6097337006427915
              f1-score: 0.6349633695864275
              support: 5445
            name: Weighted Avg

SetFit with BAAI/bge-small-en-v1.5 on Health Information Needs

This is a SetFit model that can be used for Text Classification. This SetFit model uses BAAI/bge-small-en-v1.5 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: BAAI/bge-small-en-v1.5
Classification head: a LogisticRegression instance
Maximum Sequence Length: 512 tokens
Number of Classes: 2 classes
Language: en
License: apache-2.0

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
1	'Where would you put refuse if you do not want it to exist any more?' 'Which is an example of the absolutism under peter the great?' 'where does the water at niagra falls come from'
0	'the nerves,blood vessels, and glands are located in which layer of the skin' 'Of what discipline is affective computing a branch?' 'Obesity can cause resistance to which hormone?'

Evaluation

Metrics

Label	0	1	Accuracy	Macro Avg	Weighted Avg
all	{'precision': 0.36527485731450887, 'recall': 0.990228013029316, 'f1-score': 0.5336844415185429, 'support': 1228.0}	{'precision': 0.994328922495274, 'recall': 0.4989328906805786, 'f1-score': 0.6644560240012632, 'support': 4217.0}	0.6097	{'precision': 0.6798018899048914, 'recall': 0.7445804518549473, 'f1-score': 0.5990702327599031, 'support': 5445.0}	{'precision': 0.8524596126620364, 'recall': 0.6097337006427915, 'f1-score': 0.6349633695864275, 'support': 5445.0}

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("Referees")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	1	7.3	15

Label	Training Sample Count
0	5
1	5

Training Hyperparameters

batch_size: (32, 32)
num_epochs: (10, 10)
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: False

Training Results

Epoch	Step	Training Loss	Validation Loss
0.5	1	0.1884	-

Framework Versions

Python: 3.12.2
SetFit: 1.1.0
Sentence Transformers: 3.0.1
Transformers: 4.45.2
PyTorch: 2.2.2
Datasets: 3.1.0
Tokenizers: 0.20.3

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}
}