oneryalcin/rvl-cdip-setfit-neomme

SetFit with Hcompany/NeoMME-260M-Retriever-ST-dense

This is a SetFit model that can be used for Text Classification. This SetFit model uses Hcompany/NeoMME-260M-Retriever-ST-dense 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:

1. 1.Fine-tuning a Sentence Transformer with contrastive learning.
2. 2.Training a classification head with features from the fine-tuned Sentence Transformer.

Model Details

Model Description

• —Model Type: SetFit
• —Sentence Transformer body: Hcompany/NeoMME-260M-Retriever-ST-dense
• —Classification head: a LogisticRegression instance
• —Maximum Sequence Length: 16384 tokens
• —Number of Classes: 16 classes <!-- - Training Dataset: Unknown --> <!-- - Language: Unknown --> <!-- - License: Unknown -->

Model Sources

• —Repository: SetFit on GitHub
• —Paper: Efficient Few-Shot Learning Without Prompts
• —Blogpost: SetFit: Efficient Few-Shot Learning Without Prompts

Model Labels

Uses

Direct Use for Inference

First install the SetFit library:

pip install setfit

Then you can load this model and run inference.

from PIL import Image

from setfit import SetFitModel

# Download from the 🤗 Hub
model = SetFitModel.from_pretrained("oneryalcin/rvl-cdip-setfit-neomme")
# Run inference on images
preds = model([Image.open("example.png")])

<!--

Downstream Use

List how someone could finetune this model on their own dataset. -->

<!--

Out-of-Scope Use

List how the model may foreseeably be misused and address what users ought not to do with the model. -->

<!--

Bias, Risks and Limitations

What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. -->

<!--

Recommendations

What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. -->

Training Details

Training Hyperparameters

• —batch_size: (4, 4)
• —num_epochs: (1, 1)
• —max_steps: -1
• —sampling_strategy: oversampling
• —num_iterations: 5
• —bodylearningrate: (2e-05, 1e-05)
• —headlearningrate: 0.01
• —loss: CosineSimilarityLoss
• —distancemetric: cosinedistance
• —margin: 0.25
• —endtoend: False
• —use_amp: False
• —warmup_proportion: 0.1
• —l2_weight: 0.01
• —seed: 42
• —evalmaxsteps: -1
• —loadbestmodelatend: False

Training Results

Framework Versions

• —Python: 3.12.12
• —SetFit: 1.3.0.dev0
• —Sentence Transformers: 6.0.1
• —Transformers: 5.17.0
• —PyTorch: 2.14.0+cu130
• —Datasets: 5.0.1
• —Tokenizers: 0.23.2

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}
}

<!--

Glossary

Clearly define terms in order to be accessible across audiences. -->

<!--

Model Card Authors

Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. -->

<!--

Model Card Contact

Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->

Evaluation (jordyvl/rvlcdip100examplesper_class, split test)

accuracy 0.532, macro F1 0.516 on 400 images; 8 training images per class; body Hcompany/NeoMME-260M-Retriever-ST-dense, task document; trained in 769s on cuda.

                        precision    recall  f1-score   support

         advertisement       0.56      0.76      0.64        25
                budget       0.41      0.28      0.33        25
                 email       0.64      0.56      0.60        25
           file folder       0.54      0.76      0.63        25
                  form       0.41      0.48      0.44        25
           handwritten       0.83      0.76      0.79        25
               invoice       0.33      0.24      0.28        25
                letter       0.58      0.60      0.59        25
                  memo       0.30      0.32      0.31        25
          news article       0.50      0.64      0.56        25
          presentation       0.33      0.24      0.28        25
         questionnaire       0.38      0.36      0.37        25
                resume       1.00      0.92      0.96        25
scientific publication       0.61      0.76      0.68        25
     scientific report       0.43      0.12      0.19        25
         specification       0.53      0.72      0.61        25

              accuracy                           0.53       400
             macro avg       0.52      0.53      0.52       400
          weighted avg       0.52      0.53      0.52       400