---
tags:
- setfit
- sentence-transformers
- text-classification
- generated_from_setfit_trainer
widget: []
metrics:
- accuracy
pipeline_tag: text-classification
library_name: setfit
inference: true
base_model: Hcompany/NeoMME-260M-Retriever-ST-dense
---

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

This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [Hcompany/NeoMME-260M-Retriever-ST-dense](https://huggingface.co/Hcompany/NeoMME-260M-Retriever-ST-dense) as the Sentence Transformer embedding model. A [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) 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](https://www.sbert.net) with contrastive learning.
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](https://huggingface.co/Hcompany/NeoMME-260M-Retriever-ST-dense)
- **Classification head:** a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance
- **Maximum Sequence Length:** 16384 tokens
- **Number of Classes:** 6 classes
<!-- - **Training Dataset:** [Unknown](https://huggingface.co/datasets/unknown) -->
<!-- - **Language:** Unknown -->
<!-- - **License:** Unknown -->

### Model Sources

- **Repository:** [SetFit on GitHub](https://github.com/huggingface/setfit)
- **Paper:** [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055)
- **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit)

### Model Labels
| Label                 | Examples                                                                                                                                                                                                                                                                          |
|:----------------------|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| instruction_reference | <ul><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=619x800 at 0x2ADC0091E4B0></li><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=619x800 at 0x2ADC0091F620></li><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=619x800 at 0x2ADC0091F620></li></ul> |
| news_promotion        | <ul><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=619x800 at 0x2ADC0091FB90></li><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=800x566 at 0x2ADC0091CE90></li><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=570x800 at 0x2ADC0091E630></li></ul> |
| exercise_assessment   | <ul><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=566x800 at 0x2ADC0091E180></li><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=619x800 at 0x2ADC0091FB90></li><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=566x800 at 0x2ADC0091E630></li></ul> |
| administration_policy | <ul><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=619x800 at 0x2ADC0091E750></li><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=566x800 at 0x2ADC0091F7D0></li><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=619x800 at 0x2ADC0091E4B0></li></ul> |
| research_analysis     | <ul><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=619x800 at 0x2ADC0091CCE0></li><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=619x800 at 0x2ADC0091F7D0></li><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=608x800 at 0x2ADC0091CCE0></li></ul> |
| other                 | <ul><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=619x800 at 0x2ADC0091E630></li><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=619x800 at 0x2ADC0091D5E0></li><li><PIL.PngImagePlugin.PngImageFile image mode=RGB size=619x800 at 0x2ADC0091D5E0></li></ul> |

## Uses

### Direct Use for Inference

First install the SetFit library:

```bash
pip install setfit
```

Then you can load this model and run inference.

```python
from PIL import Image

from setfit import SetFitModel

# Download from the 🤗 Hub
model = SetFitModel.from_pretrained("oneryalcin/finepdfs-purpose-setfit-neomme-v2")
# 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: (32, 32)
- num_epochs: (1, 1)
- max_steps: -1
- sampling_strategy: oversampling
- num_iterations: 3
- 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 |
|:------:|:----:|:-------------:|:---------------:|
| 0.0014 | 1    | 0.2760        | -               |
| 0.0686 | 50   | 0.2481        | -               |
| 0.1372 | 100  | 0.2100        | -               |
| 0.2058 | 150  | 0.1900        | -               |
| 0.2743 | 200  | 0.1748        | -               |
| 0.3429 | 250  | 0.1622        | -               |
| 0.4115 | 300  | 0.1531        | -               |
| 0.4801 | 350  | 0.1388        | -               |
| 0.5487 | 400  | 0.1345        | -               |
| 0.6173 | 450  | 0.1241        | -               |
| 0.6859 | 500  | 0.1158        | -               |
| 0.7545 | 550  | 0.1059        | -               |
| 0.8230 | 600  | 0.0968        | -               |
| 0.8916 | 650  | 0.0840        | -               |
| 0.9602 | 700  | 0.0874        | -               |

### Framework Versions
- Python: 3.12.10
- 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
```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 (oneryalcin/finepdfs-purpose-pages-v2, split `test`)

accuracy 0.681, macro F1 0.638 on 702 images; 800 training images per class; body `Hcompany/NeoMME-260M-Retriever-ST-dense`, task `document`; trained in 1056s on cuda.

```
                       precision    recall  f1-score   support

administration_policy       0.59      0.70      0.64        67
  exercise_assessment       0.71      0.81      0.76       118
instruction_reference       0.66      0.61      0.63       220
       news_promotion       0.63      0.70      0.66       132
                other       0.47      0.29      0.36        24
    research_analysis       0.83      0.72      0.77       141

             accuracy                           0.68       702
            macro avg       0.65      0.64      0.64       702
         weighted avg       0.68      0.68      0.68       702
```