--- library_name: setfit tags: - setfit - sentence-transformers - text-classification - generated_from_setfit_trainer metrics: - accuracy widget: - text: Hi Jonathan, I hope you're having safe travels along your way. I'm reaching out to you because you are a valued employee, and we appreciate your hard work and research. While I understand you are passionate about these projects, it is imperative that you keep your reports concise, seeing as we are all continuously on a time crunch. Because these reports are not written as efficiently as possible, it is taking too much of our time to read and determine which bit of information is most valuable. I need you to shift the way you are writing these reports so that way we can maximize our work flow processes. We love having you on our team, but if you can not make these necessary changes, we may have to relocate your skill set to a different department. However, I am positive you can make these minor changes in the way you create your reports. Please research the formal way to write reports so that way you no longer add too much information. These reports should have less opinions, and more facts. I will also send some material for you to review on how to keep these reports business friendly. I love your passion and your drive, I am hoping we can continue to have you on this project. A few minor changes will be all it takes to get the ball rolling in the right direction! If you have any concerns, feel free to reach out to me and I will be more than happy to assist. Thank you, William - text: 'Hi Jonathan, I have been hearing about some of the great work you''re doing on the Beta project, and wanted to touch base with you on how things are progressing, and what more we can do together to help you perform even better than what you are already doing Jonathan, Terry has been happy with your work on this project and even mentioned to me that you have been able to find improvements we didn''t know we needed, but as we move ahead, the team has a few concerns they would like us to address - a. Your reports with the technical information have your perspectives on the findings, not the technical information itself - we need to address this topic b. You need to improve your business writing skills in order to take the next leapI know you have been working very hard on this and your performance speaks for it, and I know your ambition to become even better, and in that spirit, let''s focus on how you can address the above mentioned issues. You are a great asset, and that''s why I need you to commit to a development plan in order for us to ensure you function at the highest level.We need to commit to the following plan of action: a. You start by preparing the technical report only with findings, not your perspectives. We value your insights, and would love to have them, but in a short memo on top of the technical report to summarize. b. We need to coach you by getting you into a business writing course - you''re a great technical engineer, but in order to rise up the ladders in business, this is an essential skill that you need to gain. I would like to hear your side of the story: your view on generating insights, what are the things we can help you out with : are there any problems you are having with the team, what extra coaching we can provide, what are your ambitions...' - text: Hi Jonathan, I would like to bring to your attention that your report writing should be improved. Your contribution and fact gathering are highly appreciated. However, when you compose the ideas into reports, it will be more productive to the team if you could separate the facts from your opinions. Your reports influence some very critical decisions at ABC-5. So a well written report will benefit many people including having higher visibility to high-ranking managers. Please clarify with Terry on report format that is most useful for him. Please keep the promised deadline. Terry needs your report so that he can compose the project report for the higher managers. Please keep the promised deadline.Please refrain from adding opinions in the report and mixing with facts. If needed, you can add a summary or conclusion as your insight.Can I have your words that you will write a good report? Please CC me in your report to Terry in the next 4 weeks. Let me know if you have any questions or concerns. Regards, William - text: Hello Jonathan, I hope you day is going well. The purpose of this msg is to improve your communication regarding your work on the Beta Project. You are important which is why we need to make sure that your thoughts and Ideas are clearly communicated with helpful factual info. I want to get your thoughts on how you best communicate and your thoughts on how to communicate more concisely. Please come up with 2-3 suggestions as will I and lets set up a time within the next 48 hours that you and I can build a plan that will help ensure your great work is being understood for the success of Beta. I am confident that we will develop a plan that continues allow your work to help the program. Please meg me what time works best for you when you end your travel. Best, William - text: Hi Jonathan, I understand you have been quite involved with the Beta Project. Your experience is paying off as you are often finding improvements the product team did not even know they needed. I wanted to share some feedback I got from one of your colleagues regarding your reports. Your enthusiasm for this project is infectious and I love to see this level of engagement. However, we also want to be mindful of the end users of the reports you are preparing. In these projects, deadlines often move at a fast pace. In order to ensure the project can stay on time, it is important to focus on inputting mainly facts when writing these reports. You offer a unique perspective and your insights are greatly appreciated. I would love to discuss your ideas with you in separate meetings outside of this project. I understand you are having to compile and organize a large amount of information. I appreciate how overwhelming this can feel at times. When these reports are completed, they are reviewed by our CEO and other key stakeholders. To ensure we are respecting their time, we want these reports to by concise and well organized. I would like you to set up some time with Terry to go over his approach to these reports and his writing style. Once I am back from assignment I will set up time to review how this meeting went and discuss other ideas you may have. I greatly appreciate your efforts on this project and positive attitude. With the above mentioned areas of opportunity, I know this project will continue to run smoothly. Thanks. pipeline_tag: text-classification inference: true base_model: sentence-transformers/paraphrase-mpnet-base-v2 --- # SetFit with sentence-transformers/paraphrase-mpnet-base-v2 This 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. 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/paraphrase-mpnet-base-v2](https://huggingface.co/sentence-transformers/paraphrase-mpnet-base-v2) - **Classification head:** a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance - **Maximum Sequence Length:** 512 tokens - **Number of Classes:** 2 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 | | ## 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("gmenchetti/paraphrase-mpnet-base-v2-empathy") # Run inference preds = model("Hello Jonathan, I hope you day is going well. The purpose of this msg is to improve your communication regarding your work on the Beta Project. You are important which is why we need to make sure that your thoughts and Ideas are clearly communicated with helpful factual info. I want to get your thoughts on how you best communicate and your thoughts on how to communicate more concisely. Please come up with 2-3 suggestions as will I and lets set up a time within the next 48 hours that you and I can build a plan that will help ensure your great work is being understood for the success of Beta. I am confident that we will develop a plan that continues allow your work to help the program. Please meg me what time works best for you when you end your travel. Best, William") ``` ## Training Details ### Training Set Metrics | Training set | Min | Median | Max | |:-------------|:----|:---------|:----| | Word count | 95 | 213.2333 | 377 | | Label | Training Sample Count | |:------|:----------------------| | 0 | 13 | | 1 | 17 | ### Training Hyperparameters - batch_size: (4, 4) - num_epochs: (3, 3) - max_steps: -1 - sampling_strategy: oversampling - num_iterations: 20 - body_learning_rate: (2e-05, 2e-05) - head_learning_rate: 2e-05 - 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 | |:------:|:----:|:-------------:|:---------------:| | 0.0033 | 1 | 0.3705 | - | | 0.1667 | 50 | 0.2017 | - | | 0.3333 | 100 | 0.0503 | - | | 0.5 | 150 | 0.0006 | - | | 0.6667 | 200 | 0.0005 | - | | 0.8333 | 250 | 0.0001 | - | | 1.0 | 300 | 0.0002 | - | | 1.1667 | 350 | 0.0002 | - | | 1.3333 | 400 | 0.0001 | - | | 1.5 | 450 | 0.0001 | - | | 1.6667 | 500 | 0.0 | - | | 1.8333 | 550 | 0.0 | - | | 2.0 | 600 | 0.0001 | - | | 2.1667 | 650 | 0.0001 | - | | 2.3333 | 700 | 0.0 | - | | 2.5 | 750 | 0.0 | - | | 2.6667 | 800 | 0.0001 | - | | 2.8333 | 850 | 0.0 | - | | 3.0 | 900 | 0.0 | - | ### Framework Versions - Python: 3.10.13 - SetFit: 1.0.3 - Sentence Transformers: 2.6.1 - Transformers: 4.39.3 - PyTorch: 2.0.0.post200 - Datasets: 2.16.1 - 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} } ```