mann2107/BCMPIIRAB_MiniLM_ALLNew
07
1---2base_model: sentence-transformers/all-MiniLM-L6-v23library_name: setfit4metrics:5- accuracy6pipeline_tag: text-classification7tags:8- setfit9- sentence-transformers10- text-classification11- generated_from_setfit_trainer12widget:13- text: Thank you for your email. Please go ahead and issue. Please invoice in KES14- text: Hi, We are missing some invoices, can you please provide it. 02 - 12 - 202015 AGENT FEE 8900784339018 $21.00 02 - 19 - 2020 AGENT FEE 0017417554160 $22.00 0216 - 19 - 2020 AGENT FEE 0017417554143 $22.00 02 - 19 - 2020 AGENT FEE 890078338342017 $21.0018- text: We need your assistance with the payment for the recent office supplies order.19 Let us know once it's done.20- text: I have reported this in November and not only was the trip supposed to be21 cancelled and credited I was double billed and the billing has not been corrected.22 The total credit should be $667.20. Please confirm this will be done.23- text: The invoice for the travel arrangements needs to be settled. Kindly provide24 payment confirmation.25inference: true26---27 28# SetFit with sentence-transformers/all-MiniLM-L6-v229 30This 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.31 32The model has been trained using an efficient few-shot learning technique that involves:33 341. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning.352. Training a classification head with features from the fine-tuned Sentence Transformer.36 37## Model Details38 39### Model Description40- **Model Type:** SetFit41- **Sentence Transformer body:** [sentence-transformers/all-MiniLM-L6-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2)42- **Classification head:** a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance43- **Maximum Sequence Length:** 256 tokens44- **Number of Classes:** 14 classes45<!-- - **Training Dataset:** [Unknown](https://huggingface.co/datasets/unknown) -->46<!-- - **Language:** Unknown -->47<!-- - **License:** Unknown -->48 49### Model Sources50 51- **Repository:** [SetFit on GitHub](https://github.com/huggingface/setfit)52- **Paper:** [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055)53- **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit)54 55 56## Uses57 58### Direct Use for Inference59 60First install the SetFit library:61 62```bash63pip install setfit64```65 66Then you can load this model and run inference.67 68```python69from setfit import SetFitModel70 71# Download from the 🤗 Hub72model = SetFitModel.from_pretrained("mann2107/BCMPIIRAB_MiniLM_ALLNew")73# Run inference74preds = model("Thank you for your email. Please go ahead and issue. Please invoice in KES")75```76 77<!--78### Downstream Use79 80*List how someone could finetune this model on their own dataset.*81-->82 83<!--84### Out-of-Scope Use85 86*List how the model may foreseeably be misused and address what users ought not to do with the model.*87-->88 89<!--90## Bias, Risks and Limitations91 92*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*93-->94 95<!--96### Recommendations97 98*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*99-->100 101## Training Details102 103### Training Set Metrics104| Training set | Min | Median | Max |105|:-------------|:----|:--------|:----|106| Word count | 1 | 25.6577 | 136 |107 108| Label | Training Sample Count |109|:------|:----------------------|110| 0 | 24 |111| 1 | 24 |112| 2 | 24 |113| 3 | 24 |114| 4 | 24 |115| 5 | 24 |116| 6 | 24 |117| 7 | 24 |118| 8 | 24 |119| 9 | 24 |120| 10 | 24 |121| 11 | 24 |122| 12 | 24 |123| 13 | 24 |124 125### Training Hyperparameters126- batch_size: (16, 16)127- num_epochs: (1, 1)128- max_steps: -1129- sampling_strategy: oversampling130- num_iterations: 99131- body_learning_rate: (0.0002733656643765287, 0.0002733656643765287)132- head_learning_rate: 2.7029049129688732e-05133- loss: CosineSimilarityLoss134- distance_metric: cosine_distance135- margin: 0.25136- end_to_end: False137- use_amp: False138- warmup_proportion: 0.1139- max_length: 512140- seed: 42141- eval_max_steps: -1142- load_best_model_at_end: True143 144### Training Results145| Epoch | Step | Training Loss | Validation Loss |146|:-------:|:--------:|:-------------:|:---------------:|147| 0.0002 | 1 | 0.2546 | - |148| 0.0120 | 50 | 0.1667 | - |149| 0.0241 | 100 | 0.1165 | - |150| 0.0361 | 150 | 0.0799 | - |151| 0.0481 | 200 | 0.0212 | - |152| 0.0601 | 250 | 0.0188 | - |153| 0.0722 | 300 | 0.0531 | - |154| 0.0842 | 350 | 0.0273 | - |155| 0.0962 | 400 | 0.0111 | - |156| 0.1082 | 450 | 0.0203 | - |157| 0.1203 | 500 | 0.0397 | - |158| 0.1323 | 550 | 0.0164 | - |159| 0.1443 | 600 | 0.0045 | - |160| 0.1563 | 650 | 0.0032 | - |161| 0.1684 | 700 | 0.001 | - |162| 0.1804 | 750 | 0.0011 | - |163| 0.1924 | 800 | 0.0004 | - |164| 0.2044 | 850 | 0.0009 | - |165| 0.2165 | 900 | 0.0006 | - |166| 0.2285 | 950 | 0.0008 | - |167| 0.2405 | 1000 | 0.0004 | - |168| 0.2525 | 1050 | 0.0008 | - |169| 0.2646 | 1100 | 0.0005 | - |170| 0.2766 | 1150 | 0.0006 | - |171| 0.2886 | 1200 | 0.0007 | - |172| 0.3006 | 1250 | 0.0043 | - |173| 0.3127 | 1300 | 0.0004 | - |174| 0.3247 | 1350 | 0.0005 | - |175| 0.3367 | 1400 | 0.0005 | - |176| 0.3487 | 1450 | 0.0004 | - |177| 0.3608 | 1500 | 0.0004 | - |178| 0.3728 | 1550 | 0.0005 | - |179| 0.3848 | 1600 | 0.0007 | - |180| 0.3968 | 1650 | 0.0006 | - |181| 0.4089 | 1700 | 0.0002 | - |182| 0.4209 | 1750 | 0.0006 | - |183| 0.4329 | 1800 | 0.0008 | - |184| 0.4449 | 1850 | 0.0003 | - |185| 0.4570 | 1900 | 0.0005 | - |186| 0.4690 | 1950 | 0.0003 | - |187| 0.4810 | 2000 | 0.0003 | - |188| 0.4930 | 2050 | 0.0003 | - |189| 0.5051 | 2100 | 0.0006 | - |190| 0.5171 | 2150 | 0.0003 | - |191| 0.5291 | 2200 | 0.0002 | - |192| 0.5411 | 2250 | 0.0002 | - |193| 0.5532 | 2300 | 0.0002 | - |194| 0.5652 | 2350 | 0.0004 | - |195| 0.5772 | 2400 | 0.0003 | - |196| 0.5892 | 2450 | 0.0003 | - |197| 0.6013 | 2500 | 0.0002 | - |198| 0.6133 | 2550 | 0.0002 | - |199| 0.6253 | 2600 | 0.0013 | - |200| 0.6373 | 2650 | 0.0002 | - |201| 0.6494 | 2700 | 0.0007 | - |202| 0.6614 | 2750 | 0.0004 | - |203| 0.6734 | 2800 | 0.0007 | - |204| 0.6854 | 2850 | 0.0018 | - |205| 0.6975 | 2900 | 0.0002 | - |206| 0.7095 | 2950 | 0.0003 | - |207| 0.7215 | 3000 | 0.0006 | - |208| 0.7335 | 3050 | 0.0003 | - |209| 0.7456 | 3100 | 0.0002 | - |210| 0.7576 | 3150 | 0.0002 | - |211| 0.7696 | 3200 | 0.0002 | - |212| 0.7816 | 3250 | 0.0002 | - |213| 0.7937 | 3300 | 0.0002 | - |214| 0.8057 | 3350 | 0.0001 | - |215| 0.8177 | 3400 | 0.0003 | - |216| 0.8297 | 3450 | 0.0002 | - |217| 0.8418 | 3500 | 0.0002 | - |218| 0.8538 | 3550 | 0.0002 | - |219| 0.8658 | 3600 | 0.0002 | - |220| 0.8778 | 3650 | 0.0002 | - |221| 0.8899 | 3700 | 0.0002 | - |222| 0.9019 | 3750 | 0.0005 | - |223| 0.9139 | 3800 | 0.0002 | - |224| 0.9259 | 3850 | 0.0001 | - |225| 0.9380 | 3900 | 0.0004 | - |226| 0.9500 | 3950 | 0.0001 | - |227| 0.9620 | 4000 | 0.0005 | - |228| 0.9740 | 4050 | 0.0002 | - |229| 0.9861 | 4100 | 0.0002 | - |230| 0.9981 | 4150 | 0.0001 | - |231| **1.0** | **4158** | **-** | **0.0302** |232 233* The bold row denotes the saved checkpoint.234### Framework Versions235- Python: 3.10.12236- SetFit: 1.1.0.dev0237- Sentence Transformers: 3.0.1238- Transformers: 4.42.4239- PyTorch: 2.3.1+cu121240- Datasets: 2.20.0241- Tokenizers: 0.19.1242 243## Citation244 245### BibTeX246```bibtex247@article{https://doi.org/10.48550/arxiv.2209.11055,248 doi = {10.48550/ARXIV.2209.11055},249 url = {https://arxiv.org/abs/2209.11055},250 author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren},251 keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences},252 title = {Efficient Few-Shot Learning Without Prompts},253 publisher = {arXiv},254 year = {2022},255 copyright = {Creative Commons Attribution 4.0 International}256}257```258 259<!--260## Glossary261 262*Clearly define terms in order to be accessible across audiences.*263-->264 265<!--266## Model Card Authors267 268*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*269-->270 271<!--272## Model Card Contact273 274*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*275-->