--- tags: - setfit - sentence-transformers - text-classification - generated_from_setfit_trainer widget: - text: 통화 중에 전화가 자꾸 끊기거나 기다리게 하는 기능이 있으면 좋겠는데 어떻게 해야 하나요? - text: 새로 산 휴대폰 언제부터 사용할 수 있는지 알려줄 수 있어? - text: 문자 확인하려고 몇 번을 기다려도 아무것도 안 오네요 - text: 어제 핸드폰이 갑자기 사라져서 당황스러워요 어디서 잃어버린 건지 모르겠네요 - text: 요즘 이상한 번호에서 자꾸 전화가 와서 좀 막을 방법이 있을까요? metrics: - accuracy pipeline_tag: text-classification library_name: setfit inference: true base_model: sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 --- # SetFit with sentence-transformers/paraphrase-multilingual-MiniLM-L12-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-multilingual-MiniLM-L12-v2](https://huggingface.co/sentence-transformers/paraphrase-multilingual-MiniLM-L12-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-multilingual-MiniLM-L12-v2](https://huggingface.co/sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2) - **Classification head:** a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance - **Maximum Sequence Length:** 128 tokens - **Number of Classes:** 36 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 | |:---------------------------|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | [단말기]모바일 U+Shop | | | [콜봇공통]장애처리 | | | [콜봇상담]IPTV장애 | | | [콜봇상담]결합상품문의 | | | [콜봇상담]기기변경문의 | | | [콜봇상담]납부방법변경 | | | [콜봇상담]납부확인서 발급 | | | [콜봇상담]듀얼넘버 문의 | | | [콜봇상담]모바일 부가서비스 가입 및 해지 | | | [콜봇상담]선택약정할인상태 안내 및 등록기능 | | | [콜봇상담]세금계산서발행 | | | [콜봇상담]스팸차단 서비스 신청 및 해지 | | | [콜봇상담]약정문의(공통) | | | [콜봇상담]연체문의(공통) | | | [콜봇상담]요금납부 | | | [콜봇상담]요금문의(공통) | | | [콜봇상담]요금제변경 | | | [콜봇상담]유심 구매 및 이동 문의 | | | [콜봇상담]이전설치 | | | [콜봇상담]인터넷 해지 | | | [콜봇상담]인터넷장애 | | | [콜봇상담]일반상담(공통) | | | [콜봇상담]일시정지 및 일시정지 해제(공통) | | | [콜봇상담]일시정지 및 일시정지 해제(모) | | | [콜봇상담]청구요금조회 | | | [콜봇상담]통화연결음 가입 및 해지 | | | [콜봇상담]통화중대기 가입 및 해지 | | | [콜봇상담]해지(공통) | | | [콜봇상담]홈서비스 가입 | | | [콜봇상담]환불_이중납부 | | | [콜봇상담]휴대폰 분실문의 | | | [콜봇상담]휴대폰결제 한도변경 | | | [콜봇상담]휴대폰결제(공통) | | | [콜봇상담]휴대폰보험문의및보상신청 | | | [콜봇이벤트]로밍상담 재질의 | | | [프리미어요금제약정할인]프리미어 요금제 약정할인 | | ## 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("huiyeong/setfit-callbot-synthetic-neg") # Run inference preds = model("문자 확인하려고 몇 번을 기다려도 아무것도 안 오네요") ``` ## Training Details ### Training Set Metrics | Training set | Min | Median | Max | |:-------------|:----|:--------|:----| | Word count | 5 | 10.2889 | 19 | | Label | Training Sample Count | |:---------------------------|:----------------------| | [단말기]모바일 U+Shop | 59 | | [콜봇공통]장애처리 | 58 | | [콜봇상담]IPTV장애 | 66 | | [콜봇상담]결합상품문의 | 60 | | [콜봇상담]기기변경문의 | 61 | | [콜봇상담]납부방법변경 | 58 | | [콜봇상담]납부확인서 발급 | 62 | | [콜봇상담]듀얼넘버 문의 | 57 | | [콜봇상담]모바일 부가서비스 가입 및 해지 | 62 | | [콜봇상담]선택약정할인상태 안내 및 등록기능 | 63 | | [콜봇상담]세금계산서발행 | 63 | | [콜봇상담]스팸차단 서비스 신청 및 해지 | 56 | | [콜봇상담]약정문의(공통) | 59 | | [콜봇상담]연체문의(공통) | 57 | | [콜봇상담]요금납부 | 58 | | [콜봇상담]요금문의(공통) | 60 | | [콜봇상담]요금제변경 | 57 | | [콜봇상담]유심 구매 및 이동 문의 | 59 | | [콜봇상담]이전설치 | 61 | | [콜봇상담]인터넷 해지 | 58 | | [콜봇상담]인터넷장애 | 57 | | [콜봇상담]일반상담(공통) | 59 | | [콜봇상담]일시정지 및 일시정지 해제(공통) | 63 | | [콜봇상담]일시정지 및 일시정지 해제(모) | 57 | | [콜봇상담]청구요금조회 | 61 | | [콜봇상담]통화연결음 가입 및 해지 | 55 | | [콜봇상담]통화중대기 가입 및 해지 | 64 | | [콜봇상담]해지(공통) | 61 | | [콜봇상담]홈서비스 가입 | 59 | | [콜봇상담]환불_이중납부 | 61 | | [콜봇상담]휴대폰 분실문의 | 64 | | [콜봇상담]휴대폰결제 한도변경 | 63 | | [콜봇상담]휴대폰결제(공통) | 56 | | [콜봇상담]휴대폰보험문의및보상신청 | 60 | | [콜봇이벤트]로밍상담 재질의 | 60 | | [프리미어요금제약정할인]프리미어 요금제 약정할인 | 66 | ### Training Hyperparameters - batch_size: (16, 16) - num_epochs: (5, 5) - 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 - l2_weight: 0.01 - seed: 42 - eval_max_steps: -1 - load_best_model_at_end: False ### Training Results | Epoch | Step | Training Loss | Validation Loss | |:------:|:----:|:-------------:|:---------------:| | 0.0007 | 1 | 0.2346 | - | | 0.0370 | 50 | 0.1917 | - | | 0.0741 | 100 | 0.1791 | - | | 0.1111 | 150 | 0.1577 | - | | 0.1481 | 200 | 0.1438 | - | | 0.1852 | 250 | 0.1441 | - | | 0.2222 | 300 | 0.1344 | - | | 0.2593 | 350 | 0.1236 | - | | 0.2963 | 400 | 0.1137 | - | | 0.3333 | 450 | 0.1068 | - | | 0.3704 | 500 | 0.0986 | - | | 0.4074 | 550 | 0.0878 | - | | 0.4444 | 600 | 0.0949 | - | | 0.4815 | 650 | 0.0866 | - | | 0.5185 | 700 | 0.0896 | - | | 0.5556 | 750 | 0.0835 | - | | 0.5926 | 800 | 0.0734 | - | | 0.6296 | 850 | 0.073 | - | | 0.6667 | 900 | 0.0673 | - | | 0.7037 | 950 | 0.0663 | - | | 0.7407 | 1000 | 0.0516 | - | | 0.7778 | 1050 | 0.0639 | - | | 0.8148 | 1100 | 0.0563 | - | | 0.8519 | 1150 | 0.0554 | - | | 0.8889 | 1200 | 0.0586 | - | | 0.9259 | 1250 | 0.0624 | - | | 0.9630 | 1300 | 0.0598 | - | | 1.0 | 1350 | 0.0494 | - | | 1.0370 | 1400 | 0.0475 | - | | 1.0741 | 1450 | 0.0424 | - | | 1.1111 | 1500 | 0.0439 | - | | 1.1481 | 1550 | 0.0386 | - | | 1.1852 | 1600 | 0.0386 | - | | 1.2222 | 1650 | 0.0415 | - | | 1.2593 | 1700 | 0.0371 | - | | 1.2963 | 1750 | 0.039 | - | | 1.3333 | 1800 | 0.0382 | - | | 1.3704 | 1850 | 0.037 | - | | 1.4074 | 1900 | 0.0453 | - | | 1.4444 | 1950 | 0.0292 | - | | 1.4815 | 2000 | 0.0391 | - | | 1.5185 | 2050 | 0.0266 | - | | 1.5556 | 2100 | 0.0397 | - | | 1.5926 | 2150 | 0.0289 | - | | 1.6296 | 2200 | 0.0306 | - | | 1.6667 | 2250 | 0.0247 | - | | 1.7037 | 2300 | 0.0258 | - | | 1.7407 | 2350 | 0.0227 | - | | 1.7778 | 2400 | 0.0265 | - | | 1.8148 | 2450 | 0.0274 | - | | 1.8519 | 2500 | 0.0206 | - | | 1.8889 | 2550 | 0.0191 | - | | 1.9259 | 2600 | 0.0266 | - | | 1.9630 | 2650 | 0.0282 | - | | 2.0 | 2700 | 0.0286 | - | | 2.0370 | 2750 | 0.0325 | - | | 2.0741 | 2800 | 0.0224 | - | | 2.1111 | 2850 | 0.027 | - | | 2.1481 | 2900 | 0.0129 | - | | 2.1852 | 2950 | 0.0205 | - | | 2.2222 | 3000 | 0.0194 | - | | 2.2593 | 3050 | 0.0203 | - | | 2.2963 | 3100 | 0.0186 | - | | 2.3333 | 3150 | 0.0145 | - | | 2.3704 | 3200 | 0.0197 | - | | 2.4074 | 3250 | 0.0244 | - | | 2.4444 | 3300 | 0.0144 | - | | 2.4815 | 3350 | 0.0139 | - | | 2.5185 | 3400 | 0.0182 | - | | 2.5556 | 3450 | 0.0203 | - | | 2.5926 | 3500 | 0.0217 | - | | 2.6296 | 3550 | 0.0131 | - | | 2.6667 | 3600 | 0.0296 | - | | 2.7037 | 3650 | 0.0192 | - | | 2.7407 | 3700 | 0.0183 | - | | 2.7778 | 3750 | 0.0153 | - | | 2.8148 | 3800 | 0.0182 | - | | 2.8519 | 3850 | 0.0194 | - | | 2.8889 | 3900 | 0.0176 | - | | 2.9259 | 3950 | 0.0202 | - | | 2.9630 | 4000 | 0.0211 | - | | 3.0 | 4050 | 0.0172 | - | | 3.0370 | 4100 | 0.0184 | - | | 3.0741 | 4150 | 0.016 | - | | 3.1111 | 4200 | 0.0081 | - | | 3.1481 | 4250 | 0.0175 | - | | 3.1852 | 4300 | 0.0131 | - | | 3.2222 | 4350 | 0.014 | - | | 3.2593 | 4400 | 0.0131 | - | | 3.2963 | 4450 | 0.0199 | - | | 3.3333 | 4500 | 0.0131 | - | | 3.3704 | 4550 | 0.0144 | - | | 3.4074 | 4600 | 0.0149 | - | | 3.4444 | 4650 | 0.0129 | - | | 3.4815 | 4700 | 0.0137 | - | | 3.5185 | 4750 | 0.0216 | - | | 3.5556 | 4800 | 0.0128 | - | | 3.5926 | 4850 | 0.0147 | - | | 3.6296 | 4900 | 0.0087 | - | | 3.6667 | 4950 | 0.0182 | - | | 3.7037 | 5000 | 0.0119 | - | | 3.7407 | 5050 | 0.0168 | - | | 3.7778 | 5100 | 0.0128 | - | | 3.8148 | 5150 | 0.0105 | - | | 3.8519 | 5200 | 0.0181 | - | | 3.8889 | 5250 | 0.0176 | - | | 3.9259 | 5300 | 0.0202 | - | | 3.9630 | 5350 | 0.0142 | - | | 4.0 | 5400 | 0.0113 | - | | 4.0370 | 5450 | 0.0105 | - | | 4.0741 | 5500 | 0.0112 | - | | 4.1111 | 5550 | 0.0133 | - | | 4.1481 | 5600 | 0.0086 | - | | 4.1852 | 5650 | 0.0167 | - | | 4.2222 | 5700 | 0.0118 | - | | 4.2593 | 5750 | 0.0096 | - | | 4.2963 | 5800 | 0.0092 | - | | 4.3333 | 5850 | 0.0096 | - | | 4.3704 | 5900 | 0.0082 | - | | 4.4074 | 5950 | 0.0143 | - | | 4.4444 | 6000 | 0.012 | - | | 4.4815 | 6050 | 0.0167 | - | | 4.5185 | 6100 | 0.0133 | - | | 4.5556 | 6150 | 0.0091 | - | | 4.5926 | 6200 | 0.0107 | - | | 4.6296 | 6250 | 0.0141 | - | | 4.6667 | 6300 | 0.0201 | - | | 4.7037 | 6350 | 0.0125 | - | | 4.7407 | 6400 | 0.0112 | - | | 4.7778 | 6450 | 0.0117 | - | | 4.8148 | 6500 | 0.0166 | - | | 4.8519 | 6550 | 0.0159 | - | | 4.8889 | 6600 | 0.0084 | - | | 4.9259 | 6650 | 0.012 | - | | 4.9630 | 6700 | 0.0098 | - | | 5.0 | 6750 | 0.0088 | - | ### Framework Versions - Python: 3.11.13 - SetFit: 1.1.2 - Sentence Transformers: 4.1.0 - Transformers: 4.52.4 - PyTorch: 2.6.0+cu124 - Datasets: 3.6.0 - Tokenizers: 0.21.2 ## 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} } ```