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 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 Sources

Model Labels

Label Examples
4
  • 'One writer , signing his letter as Red-blooded , balanced male , remarked on the frequency of women fainting in peals , and suggested that they settle back into their traditional role of making tea at meetings . '
  • 'No offense intended , he said gently . '
  • "It 's my line of work , he said "
3
  • "It was the most exercise we 'd had all morning and it was followed by our driving immediately to the nearest watering hole . "
  • 'Alimta is used together with cisplatin ( another anticancer medicine ) when the cancer is unresectable ( cannot be removed by surgery alone ) and malignant ( has spread , or is likely to spread easily , to other parts of the body ) , in patients who have not received chemotherapy ( medicines for cancer ) before advanced or metastatic non-small cell lung cancer that is not affecting the squamous cells . '
  • 'If it is , it will be treated as an operator , if it is not , it will be treated as a user function . '
6
  • '3 -RRB- Republican congressional representatives , because of their belief in a minimalist state , are less willing to engage in local benefit-seeking than are Democratic members of Congress . '
  • 'The idea would be to administer to patients the growth-controlling proteins made by healthy versions of the damaged genes . '
  • 'That is the way the system works . '
0
  • 'Prior to 1932 , the pattern was nearly the opposite . '
  • 'Never in my life have I been so frightened . '
  • 'Then your focus will go to an input text box where you can type your function . '
1
  • 'Mr. Neuberger realized that , although of Italian ancestry , Mr. Mariotta still could qualify as a minority person since he was born in Puerto Rico . '
  • 'But Dr. Vogelstein had yet to nail the identity of the gene that , if damaged , flipped a colon cell into full-blown malignancy . '
  • 'Some found it on the screen of a personal computer . '
5
  • "On the Right , the tone was set by Jacques Chirac , who declared in 1976 that 900,000 unemployed would not become a problem in a country with 2 million of foreign workers , '' and on the Left by Michel Rocard explaining in 1990 that France can not accommodate all the world 's misery . '' "
  • "But the council 's program to attract and train ringers is only partly successful , says Mr. Baldwin . "
  • 'The scientists say that since breast cancer often strikes multiple members of certain families , the gene , when inherited in a damaged form , may predispose women to the cancer . '
2
  • 'It explains how the Committee for Medicinal Products for Veterinary Use ( CVMP ) assessed the studies performed , to reach their recommendations on how to use the medicine . '
  • 'US banks repay state support '
  • '-- In most states , increasing expenditures on education , in our current circumstances , will probably make things worse , not better . '

Evaluation

Metrics

Label Accuracy
all 0.1272

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("HelgeKn/SemEval-multi-class-6")
# Run inference
preds = model("To break the uncomfortable silence , Haney began to talk . ")

Training Details

Training Set Metrics

Training set Min Median Max
Word count 4 25.0952 74
Label Training Sample Count
0 6
1 6
2 6
3 6
4 6
5 6
6 6

Training Hyperparameters

  • batch_size: (16, 16)
  • num_epochs: (2, 2)
  • max_steps: -1
  • sampling_strategy: oversampling
  • num_iterations: 20
  • body_learning_rate: (2e-05, 2e-05)
  • head_learning_rate: 2e-05
  • loss: CosineSimilarityLoss
  • distance_metric: cosine_distance
  • margin: 0.25
  • end_to_end: False
  • use_amp: False
  • warmup_proportion: 0.1
  • seed: 42
  • eval_max_steps: -1
  • load_best_model_at_end: False

Training Results

Epoch Step Training Loss Validation Loss
0.0095 1 0.3696 -
0.4762 50 0.1725 -
0.9524 100 0.0204 -
1.4286 150 0.0051 -
1.9048 200 0.0037 -

Framework Versions

  • Python: 3.9.13
  • SetFit: 1.0.1
  • Sentence Transformers: 2.2.2
  • Transformers: 4.36.0
  • PyTorch: 2.1.1+cpu
  • Datasets: 2.15.0
  • Tokenizers: 0.15.0

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
24
Safetensors
Model size
109M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for HelgeKn/SemEval-multi-class-6

Finetuned
(256)
this model

Evaluation results