metadata

library_name: setfit
tags:
  - setfit
  - sentence-transformers
  - text-classification
  - generated_from_setfit_trainer
metrics:
  - accuracy
widget:
  - text: >-
      part-of-speech ( pos ) tagging is a fundamental language analysis
      task---part-of-speech ( pos ) tagging is a fundamental nlp task , used by
      a wide variety of applications
  - text: >-
      the two baseline methods were implemented using scikit-learn in
      python---the models were implemented using scikit-learn module
  - text: >-
      semantic parsing is the task of converting a sentence into a
      representation of its meaning , usually in a logical form grounded in the
      symbols of some fixed ontology or relational database ( cite-p-21-3-3 ,
      cite-p-21-3-4 , cite-p-21-1-11 )---for this language model , we built a
      trigram language model with kneser-ney smoothing using srilm from the same
      automatically segmented corpus
  - text: >-
      the results show that our model can clearly outperform the baselines in
      terms of three evaluation metrics---for the extractive or abstractive
      summaries , we use rouge scores , a metric used to evaluate automatic
      summarization performance , to measure the pairwise agreement of summaries
      from different annotators
  - text: >-
      language models were built with srilm , modified kneser-ney smoothing ,
      default pruning , and order 5---the language model used was a 5-gram with
      modified kneserney smoothing , built with srilm toolkit
pipeline_tag: text-classification
inference: true
base_model: sentence-transformers/paraphrase-TinyBERT-L6-v2

SetFit with sentence-transformers/paraphrase-TinyBERT-L6-v2

This is a SetFit model that can be used for Text Classification. This SetFit model uses sentence-transformers/paraphrase-TinyBERT-L6-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:

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: sentence-transformers/paraphrase-TinyBERT-L6-v2
Classification head: a LogisticRegression instance
Maximum Sequence Length: 128 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
0	'the defacto standard metric in machine translation is bleu---from character representations , we propose to generate vector representations of entire tweets from characters in our tweet2vec model' 'arabic is a highly inflectional language with 85 % of words derived from trilateral roots ( alfedaghi and al-anzi 1989 )---chen et al derive bilingual subtree constraints with auto-parsed source-language sentences' 'labeling of sentence boundaries is a necessary prerequisite for many natural language processing tasks , including part-of-speech tagging and sentence alignment---we have proposed a model for video description which uses neural networks for the entire pipeline from pixels to sentences'
1	'in this paper , we present a comprehensive analysis of the relationship between personal traits and brand preferences---in previous research , in this study , we want to systematically investigate the relationship between a comprehensive set of personal traits and brand preferences' 'the 50-dimensional pre-trained word embeddings are provided by glove , which are fixed during our model training---we use glove vectors with 200 dimensions as pre-trained word embeddings , which are tuned during training' 'we use glove vectors with 100 dimensions trained on wikipedia and gigaword as word embeddings , which we do not optimize during training---we use glove vectors with 100 dimensions trained on wikipedia and gigaword as word embeddings'

Label

Examples

'the defacto standard metric in machine translation is bleu---from character representations , we propose to generate vector representations of entire tweets from characters in our tweet2vec model'
'arabic is a highly inflectional language with 85 % of words derived from trilateral roots ( alfedaghi and al-anzi 1989 )---chen et al derive bilingual subtree constraints with auto-parsed source-language sentences'
'labeling of sentence boundaries is a necessary prerequisite for many natural language processing tasks , including part-of-speech tagging and sentence alignment---we have proposed a model for video description which uses neural networks for the entire pipeline from pixels to sentences'

'in this paper , we present a comprehensive analysis of the relationship between personal traits and brand preferences---in previous research , in this study , we want to systematically investigate the relationship between a comprehensive set of personal traits and brand preferences'
'the 50-dimensional pre-trained word embeddings are provided by glove , which are fixed during our model training---we use glove vectors with 200 dimensions as pre-trained word embeddings , which are tuned during training'
'we use glove vectors with 100 dimensions trained on wikipedia and gigaword as word embeddings , which we do not optimize during training---we use glove vectors with 100 dimensions trained on wikipedia and gigaword as word embeddings'

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("whateverweird17/parasci3_1")
# Run inference
preds = model("the two baseline methods were implemented using scikit-learn in python---the models were implemented using scikit-learn module")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	27	35.8125	54

Label	Training Sample Count
0	8
1	8

Training Hyperparameters

batch_size: (8, 8)
num_epochs: (10, 10)
max_steps: -1
sampling_strategy: oversampling
num_iterations: 10
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
0.025	1	0.1715	-
1.25	50	0.0028	-
2.5	100	0.0005	-
3.75	150	0.0002	-
5.0	200	0.0003	-
6.25	250	0.0001	-
7.5	300	0.0002	-
8.75	350	0.0001	-
10.0	400	0.0001	-

Framework Versions

Python: 3.10.12
SetFit: 1.0.1
Sentence Transformers: 2.2.2
Transformers: 4.33.0
PyTorch: 2.0.0
Datasets: 2.16.0
Tokenizers: 0.13.3

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}
}