--- base_model: sentence-transformers/all-MiniLM-L6-v2 library_name: setfit metrics: - accuracy pipeline_tag: text-classification tags: - setfit - sentence-transformers - text-classification - generated_from_setfit_trainer widget: - text: Despite the widespread use of genome-based methods for taxonomic classification, some researchers (e.g., Brown and Caporaso, 2012) argue that a polyphasic approach, which combines multiple lines of evidence, remains essential for accurate Actinobacteria taxonomy. - text: This study adds to the existing literature on late-onset sepsis in very low birth weight neonates by providing insights into the clinical characteristics, microbiological etiologies, and outcomes of these infections based on a large, multicenter database. - text: The model-based clustering algorithm, specifically the Gaussian Mixture Model, effectively identified distinct clusters in the data with high accuracy. - text: The study demonstrates that waste cooking oil can be effectively converted into biodiesel using the proposed process design, yielding a high-quality fuel with significant reductions in greenhouse gas emissions. - text: TopHat and Cufflinks have been shown to outperform other tools in accurately aligning RNA-seq reads and quantifying gene and transcript expression levels, respectively (Kim et al., 2013; Trapnell et al., 2012) inference: true model-index: - name: SetFit with sentence-transformers/all-MiniLM-L6-v2 results: - task: type: text-classification name: Text Classification dataset: name: Unknown type: unknown split: test metrics: - type: accuracy value: 0.47572815533980584 name: Accuracy --- # SetFit with sentence-transformers/all-MiniLM-L6-v2 This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [sentence-transformers/all-MiniLM-L6-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L6-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. 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:** [sentence-transformers/all-MiniLM-L6-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2) - **Classification head:** a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance - **Maximum Sequence Length:** 256 tokens - **Number of Classes:** 103 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 | |:-------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | Acknowledging limitation(s) whilst stating a finding or contribution | | | Advising cautious interpretation of the findings | | | Commenting on the findings | | | Commenting on the strengths of the current study | | | Comparing the result: contradicting previous findings | | | Comparing the result: supporting previous findings | | | Contrasting sources with ‘however’ for emphasis | | | Describing previously used methods | | | Describing questionnaire design | | | Describing the characteristics of the participants | | | Describing the limitations of the current study | | | Describing the process: adverbs of manner | | | Describing the process: expressing purpose with for | | | Describing the process: infinitive of purpose | | | Describing the process: sequence words | | | Describing the process: statistical procedures | | | Describing the process: typical verbs in the passive form | | | Describing the process: using + instrument | | | Describing the research design and the methods used | | | Describing what other writers do in their published work | | | Detailing specific limitations | | | Establishing the importance of the topic for the discipline | | | Establishing the importance of the topic for the discipline: time frame given | | | Establishing the importance of the topic for the world or society | | | Establishing the importance of the topic for the world or society: time frame given | | | Establising the importance of the topic as a problem to be addressed | | | Explaining keywords (also refer to Defining Terms) | | | Explaining the provenance of articles for review | | | Explaining the provenance of the participants | | | Explaining the significance of the current study | | | Explaining the significance of the findings or contribution of the study | | | General comments on the relevant literature | | | General reference to previous research or scholarship: highlighting negative outcomes | | | Giving reasons for personal interest in the research (sometimes found in the humanities, and the applied human sciences) | | | Giving reasons why a particular method was adopted | | | Giving reasons why a particular method was rejected | | | Highlighting inadequacies or weaknesses of previous studies (also refer to Being Critical) | | | Highlighting interesting or surprising results | | | Highlighting significant data in a table or chart | | | Identifying a controversy within the field of study | | | Identifying a knowledge gap in the field of study | | | Implications and/or recommendations for practice or policy | | | Indicating an expected outcome | | | Indicating an unexpected outcome | | | Indicating criteria for selection or inclusion in the study | | | Indicating methodological problems or limitations | | | Indicating missing, weak, or contradictory evidence | | | Indicating the methodology for the current research | | | Indicating the use of an established method | | | Introducing the limitations of the current study | | | Making recommendations for further research work | | | Noting implications of the findings | | | Noting the lack of or paucity of previous research | | | Offering an explanation for the findings | | | Outlining the structure of a short paper | | | Outlining the structure of a thesis or dissertation | | | Pointing out interesting or important findings | | | Previewing a chapter | | | Previous research: A historic perspective | | | Previous research: Approaches taken | | | Previous research: What has been established or proposed | | | Previous research: area investigated as the sentence object | | | Previous research: area investigated as the sentence subject | | | Previous research: highlighting negative outcomes | | | Providing background information: reference to the literature | | | Providing background information: reference to the purpose of the study | | | Reference to previous research: important studies | | | Referring back to the purpose of the paper or study | | | Referring back to the research aims or procedures | | | Referring to a single investigation in the past: investigation prominent | | | Referring to a single investigation in the past: researcher prominent | | | Referring to another writer’s idea(s) or position | | | Referring to data in a table or chart | | | Referring to important texts in the area of interest | | | Referring to previous work to establish what is already known | | | Referring to secondary sources | | | Referring to the literature to justify a method or approach | | | Reporting positive and negative reactions | | | Restating a result or one of several results | | | Setting out the research questions or hypotheses | | | Some ways of introducing quotations | | | Stating a negative result | | | Stating a positive result | | | Stating purpose of the current research with reference to gaps or issues in the literature | | | Stating the aims of the current research (note frequent use of past tense) | | | Stating the focus, aim, or argument of a short paper | | | Stating the purpose of the thesis, dissertation, or research article (note use of present tense) | | | Stating what is currently known about the topic | | | Suggesting general hypotheses | | | Suggesting implications for what is already known | | | Suggestions for future work | | | Summarising the literature review | | | Summarising the main research findings | | | Summarising the results section | | | Summarising the studies reviewed | | | Surveys and interviews: Introducing excerpts from interview data | | | Surveys and interviews: Reporting participants’ views | | | Surveys and interviews: Reporting proportions | | | Surveys and interviews: Reporting response rates | | | Surveys and interviews: Reporting themes | | | Synthesising sources: contrasting evidence or ideas | | | Synthesising sources: supporting evidence or ideas | | | Transition: moving to the next result | | ## Evaluation ### Metrics | Label | Accuracy | |:--------|:---------| | **all** | 0.4757 | ## 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 SetFitModel # Download from the 🤗 Hub model = SetFitModel.from_pretrained("Corran/SciGenSetfit24") # Run inference preds = model("The model-based clustering algorithm, specifically the Gaussian Mixture Model, effectively identified distinct clusters in the data with high accuracy.") ``` ## Training Details ### Training Set Metrics | Training set | Min | Median | Max | |:-------------|:----|:--------|:----| | Word count | 6 | 28.4192 | 62 | | Label | Training Sample Count | |:-------------------------------------------------------------------------------------------------------------------------|:----------------------| | Acknowledging limitation(s) whilst stating a finding or contribution | 50 | | Advising cautious interpretation of the findings | 50 | | Commenting on the findings | 50 | | Commenting on the strengths of the current study | 50 | | Comparing the result: contradicting previous findings | 50 | | Comparing the result: supporting previous findings | 50 | | Contrasting sources with ‘however’ for emphasis | 50 | | Describing previously used methods | 50 | | Describing questionnaire design | 50 | | Describing the characteristics of the participants | 50 | | Describing the limitations of the current study | 50 | | Describing the process: adverbs of manner | 50 | | Describing the process: expressing purpose with for | 50 | | Describing the process: infinitive of purpose | 50 | | Describing the process: sequence words | 50 | | Describing the process: statistical procedures | 50 | | Describing the process: typical verbs in the passive form | 50 | | Describing the process: using + instrument | 50 | | Describing the research design and the methods used | 50 | | Describing what other writers do in their published work | 50 | | Detailing specific limitations | 50 | | Establishing the importance of the topic for the discipline | 50 | | Establishing the importance of the topic for the discipline: time frame given | 50 | | Establishing the importance of the topic for the world or society | 50 | | Establishing the importance of the topic for the world or society: time frame given | 50 | | Establising the importance of the topic as a problem to be addressed | 50 | | Explaining keywords (also refer to Defining Terms) | 50 | | Explaining the provenance of articles for review | 50 | | Explaining the provenance of the participants | 50 | | Explaining the significance of the current study | 50 | | Explaining the significance of the findings or contribution of the study | 50 | | General comments on the relevant literature | 50 | | General reference to previous research or scholarship: highlighting negative outcomes | 50 | | Giving reasons for personal interest in the research (sometimes found in the humanities, and the applied human sciences) | 50 | | Giving reasons why a particular method was adopted | 50 | | Giving reasons why a particular method was rejected | 50 | | Highlighting inadequacies or weaknesses of previous studies (also refer to Being Critical) | 50 | | Highlighting interesting or surprising results | 50 | | Highlighting significant data in a table or chart | 50 | | Identifying a controversy within the field of study | 50 | | Identifying a knowledge gap in the field of study | 50 | | Implications and/or recommendations for practice or policy | 50 | | Indicating an expected outcome | 50 | | Indicating an unexpected outcome | 50 | | Indicating criteria for selection or inclusion in the study | 50 | | Indicating methodological problems or limitations | 50 | | Indicating missing, weak, or contradictory evidence | 50 | | Indicating the methodology for the current research | 50 | | Indicating the use of an established method | 50 | | Introducing the limitations of the current study | 50 | | Making recommendations for further research work | 50 | | Noting implications of the findings | 50 | | Noting the lack of or paucity of previous research | 50 | | Offering an explanation for the findings | 50 | | Outlining the structure of a short paper | 50 | | Outlining the structure of a thesis or dissertation | 50 | | Pointing out interesting or important findings | 50 | | Previewing a chapter | 50 | | Previous research: A historic perspective | 50 | | Previous research: Approaches taken | 50 | | Previous research: What has been established or proposed | 50 | | Previous research: area investigated as the sentence object | 50 | | Previous research: area investigated as the sentence subject | 50 | | Previous research: highlighting negative outcomes | 50 | | Providing background information: reference to the literature | 50 | | Providing background information: reference to the purpose of the study | 50 | | Reference to previous research: important studies | 50 | | Referring back to the purpose of the paper or study | 50 | | Referring back to the research aims or procedures | 50 | | Referring to a single investigation in the past: investigation prominent | 50 | | Referring to a single investigation in the past: researcher prominent | 50 | | Referring to another writer’s idea(s) or position | 50 | | Referring to data in a table or chart | 50 | | Referring to important texts in the area of interest | 50 | | Referring to previous work to establish what is already known | 50 | | Referring to secondary sources | 50 | | Referring to the literature to justify a method or approach | 50 | | Reporting positive and negative reactions | 50 | | Restating a result or one of several results | 50 | | Setting out the research questions or hypotheses | 50 | | Some ways of introducing quotations | 50 | | Stating a negative result | 50 | | Stating a positive result | 50 | | Stating purpose of the current research with reference to gaps or issues in the literature | 50 | | Stating the aims of the current research (note frequent use of past tense) | 50 | | Stating the focus, aim, or argument of a short paper | 50 | | Stating the purpose of the thesis, dissertation, or research article (note use of present tense) | 50 | | Stating what is currently known about the topic | 50 | | Suggesting general hypotheses | 50 | | Suggesting implications for what is already known | 50 | | Suggestions for future work | 50 | | Summarising the literature review | 50 | | Summarising the main research findings | 50 | | Summarising the results section | 50 | | Summarising the studies reviewed | 50 | | Surveys and interviews: Introducing excerpts from interview data | 50 | | Surveys and interviews: Reporting participants’ views | 50 | | Surveys and interviews: Reporting proportions | 50 | | Surveys and interviews: Reporting response rates | 50 | | Surveys and interviews: Reporting themes | 50 | | Synthesising sources: contrasting evidence or ideas | 50 | | Synthesising sources: supporting evidence or ideas | 50 | | Transition: moving to the next result | 50 | ### Training Hyperparameters - batch_size: (300, 300) - num_epochs: (1, 1) - max_steps: -1 - sampling_strategy: oversampling - num_iterations: 5 - 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 - seed: 42 - eval_max_steps: -1 - load_best_model_at_end: False ### Training Results | Epoch | Step | Training Loss | Validation Loss | |:------:|:----:|:-------------:|:---------------:| | 0.0058 | 1 | 0.4364 | - | | 0.2907 | 50 | 0.1895 | - | | 0.5814 | 100 | 0.1527 | - | | 0.8721 | 150 | 0.139 | - | ### Framework Versions - Python: 3.10.12 - SetFit: 1.0.3 - Sentence Transformers: 3.1.1 - Transformers: 4.39.0 - PyTorch: 2.5.1+cu121 - Datasets: 3.1.0 - 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} } ```