--- base_model: sentence-transformers/paraphrase-multilingual-mpnet-base-v2 library_name: setfit metrics: - f1 pipeline_tag: text-classification tags: - setfit - absa - sentence-transformers - text-classification - generated_from_setfit_trainer widget: - text: Il risotto al taleggio era eccellente, ma il:Il risotto al taleggio era eccellente, ma il posto era un po' affollato. - text: rimasti soddisfatti. Ottimi i casoncelli. Unico neo una porzione:Abbiamo pranzato all’aperto dopo una bellissima passeggiata per la Citta’ Alta. Quello che stupisce è sicuramente il rapporto qualità prezzo, in quanto abbiamo scelto il menu’ a prezzo fisso (15 euro) e ne siamo rimasti soddisfatti. Ottimi i casoncelli. Unico neo una porzione di coniglio...Altro - text: 'Posizione strategica su un poggio panoramico:Posizione strategica su un poggio panoramico di Bergamo alta con vista mozzafiato. L''ambiente è riscaldato visivamente da una boiserie continua molto particolare; ad accoglierci un piano con dolci a vista che dopo scopriamo essere dei sopraffini prodotti di pasticceria francese. Un entreè con le classiche...Altro' - text: bue era davvero ottima anche anatra e faraona, porzioni molto:La costata di bue era davvero ottima anche anatra e faraona, porzioni molto buone e saporite. L'ambiente suggestivo e molto curato. Buona la carta dei vini e servizio molto gentile e attento. Ci tornerò appena posso. - text: Posto insolito è un ex carcere:Posto insolito è un ex carcere riadattato. Non ha insegna e può sfuggire l'ingresso a chi non ci è mai stato. Il locale è affascinante, ben strutturato e molto accogliente. Fuori c’erano i mercatini. I piatti sono stati molto buoni, cucina locale. Personale gentile e...Altro inference: false model-index: - name: SetFit Polarity Model with sentence-transformers/paraphrase-multilingual-mpnet-base-v2 results: - task: type: text-classification name: Text Classification dataset: name: Unknown type: unknown split: test metrics: - type: f1 value: 0.841499174384361 name: F1 --- # SetFit Polarity Model with sentence-transformers/paraphrase-multilingual-mpnet-base-v2 This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Aspect Based Sentiment Analysis (ABSA). This SetFit model uses [sentence-transformers/paraphrase-multilingual-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-multilingual-mpnet-base-v2) as the Sentence Transformer embedding model. A [SetFitHead](huggingface.co/docs/setfit/reference/main#setfit.SetFitHead) instance is used for classification. In particular, this model is in charge of classifying aspect polarities. 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. This model was trained within the context of a larger system for ABSA, which looks like so: 1. Use a spaCy model to select possible aspect span candidates. 2. Use a SetFit model to filter these possible aspect span candidates. 3. **Use this SetFit model to classify the filtered aspect span candidates.** ## Model Details ### Model Description - **Model Type:** SetFit - **Sentence Transformer body:** [sentence-transformers/paraphrase-multilingual-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-multilingual-mpnet-base-v2) - **Classification head:** a [SetFitHead](huggingface.co/docs/setfit/reference/main#setfit.SetFitHead) instance - **spaCy Model:** it_core_news_lg - **SetFitABSA Aspect Model:** [setfit-absa-aspect](https://huggingface.co/setfit-absa-aspect) - **SetFitABSA Polarity Model:** [MattiaTintori/Final_polarity_Colab_It](https://huggingface.co/MattiaTintori/Final_polarity_Colab_It) - **Maximum Sequence Length:** 128 tokens - **Number of Classes:** 3 classes ### 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 | |:------|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | 1 | | | 0 | | | 2 | | ## Evaluation ### Metrics | Label | F1 | |:--------|:-------| | **all** | 0.8415 | ## 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 setfit import AbsaModel # Download from the 🤗 Hub model = AbsaModel.from_pretrained( "setfit-absa-aspect", "MattiaTintori/Final_polarity_Colab_It", ) # Run inference preds = model("The food was great, but the venue is just way too busy.") ``` ## Training Details ### Training Set Metrics | Training set | Min | Median | Max | |:-------------|:----|:--------|:----| | Word count | 14 | 42.1222 | 146 | | Label | Training Sample Count | |:------|:----------------------| | 0 | 914 | | 1 | 345 | | 2 | 148 | ### Training Hyperparameters - batch_size: (128, 32) - num_epochs: (5, 32) - max_steps: -1 - sampling_strategy: oversampling - num_iterations: 10 - body_learning_rate: (2e-05, 2e-05) - head_learning_rate: 0.04 - loss: CosineSimilarityLoss - distance_metric: cosine_distance - margin: 0.25 - end_to_end: False - use_amp: True - warmup_proportion: 0.1 - l2_weight: 0.02 - seed: 42 - eval_max_steps: -1 - load_best_model_at_end: True ### Training Results | Epoch | Step | Training Loss | Validation Loss | |:-------:|:-------:|:-------------:|:---------------:| | 0.0045 | 1 | 0.2501 | - | | 0.0455 | 10 | 0.2514 | 0.2407 | | 0.0909 | 20 | 0.2359 | 0.2252 | | 0.1364 | 30 | 0.21 | 0.2067 | | 0.1818 | 40 | 0.1984 | 0.1779 | | 0.2273 | 50 | 0.1408 | 0.1469 | | 0.2727 | 60 | 0.1246 | 0.1493 | | 0.3182 | 70 | 0.0654 | 0.1312 | | 0.3636 | 80 | 0.0546 | 0.1293 | | 0.4091 | 90 | 0.0651 | 0.1222 | | 0.4545 | 100 | 0.0374 | 0.1385 | | **0.5** | **110** | **0.0546** | **0.1214** | | 0.5455 | 120 | 0.0453 | 0.1284 | | 0.5909 | 130 | 0.0269 | 0.1241 | | 0.6364 | 140 | 0.0303 | 0.1451 | | 0.6818 | 150 | 0.0355 | 0.1299 | | 0.7273 | 160 | 0.0096 | 0.1329 | | 0.7727 | 170 | 0.0129 | 0.1411 | | 0.8182 | 180 | 0.0127 | 0.1325 | * The bold row denotes the saved checkpoint. ### Framework Versions - Python: 3.10.12 - SetFit: 1.0.3 - Sentence Transformers: 3.1.0 - spaCy: 3.7.6 - Transformers: 4.39.0 - PyTorch: 2.4.0+cu121 - Datasets: 3.0.0 - Tokenizers: 0.15.2 ## 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} } ```