hojzas/setfit-tutorial

1---2library_name: setfit3tags:4- setfit5- sentence-transformers6- text-classification7- generated_from_setfit_trainer8metrics:9- accuracy10widget:11- text: 'has the chops of a smart-aleck film school brat and the imagination of a12    big kid ... '13- text: 'that ''s truly deserving of its oscar nomination '14- text: 'instead gets ( sci-fi ) rehash '15- text: 'career-defining revelation '16- text: 'is ultimately about as inspiring as a hallmark card . '17pipeline_tag: text-classification18inference: true19co2_eq_emissions:20  emissions: 0.0598391654778262221  source: codecarbon22  training_type: fine-tuning23  on_cloud: false24  cpu_model: Intel(R) Xeon(R) Silver 4314 CPU @ 2.40GHz25  ram_total_size: 251.4916038513183626  hours_used: 0.00127base_model: sentence-transformers/paraphrase-mpnet-base-v228---29 30# SetFit with sentence-transformers/paraphrase-mpnet-base-v231 32This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) 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.33 34The model has been trained using an efficient few-shot learning technique that involves:35 361. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning.372. Training a classification head with features from the fine-tuned Sentence Transformer.38 39## Model Details40 41### Model Description42- **Model Type:** SetFit43- **Sentence Transformer body:** [sentence-transformers/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2)44- **Classification head:** a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance45- **Maximum Sequence Length:** 512 tokens46- **Number of Classes:** 2 classes47<!-- - **Training Dataset:** [Unknown](https://huggingface.co/datasets/unknown) -->48<!-- - **Language:** Unknown -->49<!-- - **License:** Unknown -->50 51### Model Sources52 53- **Repository:** [SetFit on GitHub](https://github.com/huggingface/setfit)54- **Paper:** [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055)55- **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit)56 57### Model Labels58| Label | Examples                                                                                                                                                                                                                                                                                         |59|:------|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|60| 0     | <ul><li>'stale and uninspired . '</li><li>"the film 's considered approach to its subject matter is too calm and thoughtful for agitprop , and the thinness of its characterizations makes it a failure as straight drama . ' "</li><li>"that their charm does n't do a load of good "</li></ul> |61| 1     | <ul><li>"broomfield is energized by volletta wallace 's maternal fury , her fearlessness "</li><li>'flawless '</li><li>'insightfully written , delicately performed '</li></ul>                                                                                                                  |62 63## Uses64 65### Direct Use for Inference66 67First install the SetFit library:68 69```bash70pip install setfit71```72 73Then you can load this model and run inference.74 75```python76from setfit import SetFitModel77 78# Download from the 🤗 Hub79model = SetFitModel.from_pretrained("hojzas/setfit-tutorial")80# Run inference81preds = model("career-defining revelation ")82```83 84<!--85### Downstream Use86 87*List how someone could finetune this model on their own dataset.*88-->89 90<!--91### Out-of-Scope Use92 93*List how the model may foreseeably be misused and address what users ought not to do with the model.*94-->95 96<!--97## Bias, Risks and Limitations98 99*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*100-->101 102<!--103### Recommendations104 105*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*106-->107 108## Training Details109 110### Training Set Metrics111| Training set | Min | Median  | Max |112|:-------------|:----|:--------|:----|113| Word count   | 2   | 11.4375 | 33  |114 115| Label | Training Sample Count |116|:------|:----------------------|117| 0     | 8                     |118| 1     | 8                     |119 120### Training Hyperparameters121- batch_size: (16, 16)122- num_epochs: (1, 1)123- max_steps: -1124- sampling_strategy: oversampling125- num_iterations: 20126- body_learning_rate: (2e-05, 2e-05)127- head_learning_rate: 2e-05128- loss: CosineSimilarityLoss129- distance_metric: cosine_distance130- margin: 0.25131- end_to_end: False132- use_amp: False133- warmup_proportion: 0.1134- seed: 42135- eval_max_steps: -1136- load_best_model_at_end: False137 138### Training Results139| Epoch | Step | Training Loss | Validation Loss |140|:-----:|:----:|:-------------:|:---------------:|141| 0.025 | 1    | 0.2176        | -               |142 143### Environmental Impact144Carbon emissions were measured using [CodeCarbon](https://github.com/mlco2/codecarbon).145- **Carbon Emitted**: 0.000 kg of CO2146- **Hours Used**: 0.001 hours147 148### Training Hardware149- **On Cloud**: No150- **GPU Model**: No GPU used151- **CPU Model**: Intel(R) Xeon(R) Silver 4314 CPU @ 2.40GHz152- **RAM Size**: 251.49 GB153 154### Framework Versions155- Python: 3.10.12156- SetFit: 1.0.3157- Sentence Transformers: 2.2.2158- Transformers: 4.36.1159- PyTorch: 2.1.2+cu121160- Datasets: 2.14.7161- Tokenizers: 0.15.1162 163## Citation164 165### BibTeX166```bibtex167@article{https://doi.org/10.48550/arxiv.2209.11055,168    doi = {10.48550/ARXIV.2209.11055},169    url = {https://arxiv.org/abs/2209.11055},170    author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren},171    keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences},172    title = {Efficient Few-Shot Learning Without Prompts},173    publisher = {arXiv},174    year = {2022},175    copyright = {Creative Commons Attribution 4.0 International}176}177```178 179<!--180## Glossary181 182*Clearly define terms in order to be accessible across audiences.*183-->184 185<!--186## Model Card Authors187 188*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*189-->190 191<!--192## Model Card Contact193 194*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*195-->