File size: 12,995 Bytes
6b255bf a8b0259 6b255bf a8b0259 6b255bf 1699c8d 6b255bf a8b0259 6b255bf a8b0259 6b255bf a8b0259 6b255bf a8b0259 6b255bf a8b0259 6b255bf a8b0259 6b255bf a8b0259 6b255bf a8b0259 6b255bf a8b0259 6b255bf a8b0259 6b255bf a8b0259 6b255bf a8b0259 6b255bf |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 |
---
tags:
- setfit
- sentence-transformers
- text-classification
- generated_from_setfit_trainer
widget:
- text: Etunimi Sukunimi Ruotsin kansasta yli puolet reilusti kannattaa Natoon hakemista.
Ainoana esteenä näkisin, että Ruotsin asevoimat eivät ole läheskään niin hyvällä
tolalla kuin Suomen mitä tulee varusmiehiin, reserviläisiin tai edes kalustoonkaan.
Mutta yhteisen hakemuksen kohdalla se tuskin olisi ongelma muille Nato-maille
hyväksynnän suhteen. Toinen ongelma on, että hyväksynnän tulisi olla sataprosenttinen
ja ei ole poissujettua, että Venäjä esmes vaikuttaisi yksittäiseen maahan niin,
että juuri se ei hyväksyisikään hakemusta.
- text: Etunimi Pugh nyt ymmärrän sun puolustelut asut jenkeissä.....
- text: Etunimi Sukunimi ei varmasti moni uskalla
- text: Etunimi Sukunimi Voisitko laittaa tuohon lastentappoväitteeseen mukaan jonkinlaista
faktaa. Jää muuten melko irralliseksi heitoksi. Ja etkös aiemmin korostanut, että
maa ei ole sama kuin ihmiset? No mikä on maa tai valtio, se on jäsentensä muodostama.
Nyt sitten väität, että Ukraina on maana tappanut lapsia 8 vuotta.
- text: Etunimi Sukunimi Historiaa kirjoitetaan vielä maaliskuun 2020 tapahtumista.
metrics:
- metric
pipeline_tag: text-classification
library_name: setfit
inference: true
base_model: TurkuNLP/bert-base-finnish-cased-v1
model-index:
- name: SetFit with TurkuNLP/bert-base-finnish-cased-v1
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: Unknown
type: unknown
split: test
metrics:
- type: metric
value: 0.8718253349471188
name: Metric
---
# Detect Questions in Asynchronous Conversation Comments
## SetFit with TurkuNLP/bert-base-finnish-cased-v1
This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification of actions in asynchronous conversation. This particular model detects if a comment includes a question or not. The configuration of the model is that the model is based on only one annotator's annotations (annotator A3). Metric evaluations are based on conservative ground truth (see paper). This SetFit model uses TurkuNLP/bert-base-finnish-cased-v1 as the Sentence Transformer embedding model (using word embeddings). A LogisticRegression 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:** [TurkuNLP/bert-base-finnish-cased-v1](https://huggingface.co/TurkuNLP/bert-base-finnish-cased-v1)
- **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 (contains statement - yes or no)
<!-- - **Training Dataset:** [Unknown](https://huggingface.co/datasets/unknown) -->
<!-- - **Language:** Unknown -->
<!-- - **License:** Unknown -->
### Model Sources
- **Repository:** [GitHub](https://github.com/henniina/Detecting-paired-actions)
- **Paper:** Paakki, H., Toivanen, P. and Kajava K. (2025). Implicit and Indirect: Detecting Face-threatening and Paired Actions in Asynchronous Online Conversations. Northern European Journal of Language Technology (NEJLT), 11(1), pp. 58-83.
### Model Labels
| Label | Examples |
|:------|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| 1 | <ul><li>'Etunimi Sukunimi jep. Suomalainen rokottamaton paha, ukrainalainen rokottamaton hyvä. Tää näkyy olevan nyt se mentaliteetti tällä hetkellä...'</li><li>'Etunimi Sukunimi tilastot.'</li><li>'Etunimi Sukunimi myös delta oli suurimmalle osalle myös rokottamattomille lähes oireeton, omikron kuulemma vielä lievempi👏'</li></ul> |
| 0 | <ul><li>'Etunimi Sukunimi RAutaa rajoille Suomi suureksi ja Viena vapaaksi'</li><li>'Perussuomalaiset siivoamassa keskustelupalstoja, koronakriisiavustuksien avulla? Onhan tämä nyt joku Monty Python -sketsi?'</li><li>'on se hyvä että Kiurussa ei ole miestä vaan Niskavuoren Hetaa joka pistää tuollaisen pojanklopin aisoihin viimeistään silloin kun Vapaavuori on kaltereissa johtaessaan Uuttamaata terveyspaniikkiin.'</li></ul> |
## Evaluation
### Metrics
| Label | Metric |
|:--------|:-------|
| **5-fold cross-validated F1** | 0.78 |
## 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("Finnish-actions/SetFit-FinBERT1-A3")
# Run inference
preds = model("Etunimi Sukunimi ei varmasti moni uskalla")
```
### Downstream Use
NB. This model has been trained on data coming from Finnish language asynchronous conversations under crisis related news on Facebook. This specific model has been trained to detect whether a comment includes a question or not. It reflects only one of our annotators' label interpretations, so the best use of our models (see our paper) would be to combine a set of models we provide on our Huggingface (Finnish-actions), and use a model ensemble to provide label predictions. It needs to be noted also that the model may not be well applicable outside of its empirical context, so in downstream applications, one should always conduct an evaluation of the model applicability using manually annotated data from that specific context (see our paper for annotation instructions).
### Out-of-Scope Use
Please use this model only for action detection and analysis. Uses of this model and the involved data for generative purposes (e.g. NLG) is prohibited.
## Bias, Risks and Limitations
Note that the model may produce errors. Due to the size of the training dataset, model may not generalize very well even for other novel topics within the same context. Note that model predictions should not be regarded as final judgments e.g. for online moderation purposes, but each case should also be regarded individually if using model predictions to support moderation. Also, the annotations only reflect three (though experienced) annotators' interpretations, so there might be perspectives on data intepretation that have not been taken into account here.
If model is used to support moderation on social media, we recommend that final judgments should always be left for human moderators.
<!--
### Recommendations
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
-->
## Training Details
### Training Set Metrics
| Training set | Min | Median | Max |
|:-------------|:----|:--------|:----|
| Word count | 1 | 19.6854 | 213 |
| Label | Training Sample Count |
|:------|:----------------------|
| 0 | 263 |
| 1 | 700 |
<!--
### Training Hyperparameters
- batch_size: (16, 16)
- num_epochs: (4, 4)
- max_steps: -1
- sampling_strategy: oversampling
- num_iterations: 6
- 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
- evaluation_strategy: epoch
- eval_max_steps: -1
- load_best_model_at_end: False
### Training Results
| Epoch | Step | Training Loss | Validation Loss |
|:------:|:----:|:-------------:|:---------------:|
| 0.0014 | 1 | 0.2224 | - |
| 0.0692 | 50 | 0.2676 | - |
| 0.1383 | 100 | 0.2486 | - |
| 0.2075 | 150 | 0.2208 | - |
| 0.2766 | 200 | 0.1892 | - |
| 0.3458 | 250 | 0.1509 | - |
| 0.4149 | 300 | 0.1194 | - |
| 0.4841 | 350 | 0.0745 | - |
| 0.5533 | 400 | 0.039 | - |
| 0.6224 | 450 | 0.0298 | - |
| 0.6916 | 500 | 0.01 | - |
| 0.7607 | 550 | 0.006 | - |
| 0.8299 | 600 | 0.0021 | - |
| 0.8990 | 650 | 0.0017 | - |
| 0.9682 | 700 | 0.0038 | - |
| 1.0 | 723 | - | 0.2008 |
| 1.0373 | 750 | 0.0088 | - |
| 1.1065 | 800 | 0.0041 | - |
| 1.1757 | 850 | 0.0067 | - |
| 1.2448 | 900 | 0.0041 | - |
| 1.3140 | 950 | 0.0021 | - |
| 1.3831 | 1000 | 0.0036 | - |
| 1.4523 | 1050 | 0.0036 | - |
| 1.5214 | 1100 | 0.0011 | - |
| 1.5906 | 1150 | 0.0035 | - |
| 1.6598 | 1200 | 0.0047 | - |
| 1.7289 | 1250 | 0.0005 | - |
| 1.7981 | 1300 | 0.0002 | - |
| 1.8672 | 1350 | 0.0029 | - |
| 1.9364 | 1400 | 0.0029 | - |
| 2.0 | 1446 | - | 0.2342 |
| 2.0055 | 1450 | 0.0014 | - |
| 2.0747 | 1500 | 0.0023 | - |
| 2.1438 | 1550 | 0.0022 | - |
| 2.2130 | 1600 | 0.0014 | - |
| 2.2822 | 1650 | 0.0024 | - |
| 2.3513 | 1700 | 0.0035 | - |
| 2.4205 | 1750 | 0.0014 | - |
| 2.4896 | 1800 | 0.0022 | - |
| 2.5588 | 1850 | 0.0025 | - |
| 2.6279 | 1900 | 0.0003 | - |
| 2.6971 | 1950 | 0.0042 | - |
| 2.7663 | 2000 | 0.0014 | - |
| 2.8354 | 2050 | 0.0003 | - |
| 2.9046 | 2100 | 0.0022 | - |
| 2.9737 | 2150 | 0.0031 | - |
| 3.0 | 2169 | - | 0.2224 |
| 3.0429 | 2200 | 0.0016 | - |
| 3.1120 | 2250 | 0.0014 | - |
| 3.1812 | 2300 | 0.005 | - |
| 3.2503 | 2350 | 0.0045 | - |
| 3.3195 | 2400 | 0.001 | - |
| 3.3887 | 2450 | 0.0012 | - |
| 3.4578 | 2500 | 0.0004 | - |
| 3.5270 | 2550 | 0.0013 | - |
| 3.5961 | 2600 | 0.0022 | - |
| 3.6653 | 2650 | 0.0009 | - |
| 3.7344 | 2700 | 0.0018 | - |
| 3.8036 | 2750 | 0.0015 | - |
| 3.8728 | 2800 | 0.0019 | - |
| 3.9419 | 2850 | 0.0025 | - |
| 4.0 | 2892 | - | 0.2222 |
-->
### Framework Versions
- Python: 3.11.9
- SetFit: 1.1.3
- Sentence Transformers: 3.2.0
- Transformers: 4.44.0
- PyTorch: 2.4.0+cu124
- Datasets: 2.21.0
- Tokenizers: 0.19.1
## Citation
### BibTeX
```bibtex
@article{paakki-implicit-indirect,
doi = {https://doi.org/10.3384/nejlt.2000-1533.2025.5980},
url = {https://nejlt.ep.liu.se/article/view/5980},
author = {Paakki, Henna and Toivanen, Pihla and Kajava, Kaisla},
title = {Implicit and Indirect: Detecting Face-threatening and Paired Actions in Asynchronous Online Conversations},
publisher = {Northern European Journal of Language Technology (NEJLT)},
volume= {11},
number= {1},
year = {2025}
}
```
<!--
## Glossary
*Clearly define terms in order to be accessible across audiences.*
-->
<!--
## Model Card Authors
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
-->
<!--
## Model Card Contact
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
--> |