Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Suicidal-ELECTRA
|
| 2 |
+
This text classification model predicts whether a sequence of words are suicidal (1) or non-suicidal (0).
|
| 3 |
+
|
| 4 |
+
## Data
|
| 5 |
+
The model was trained on the [Suicide and Depression Dataset](https://www.kaggle.com/nikhileswarkomati/suicide-watch) obtained from Kaggle. The dataset was scraped from Reddit and consists of 232,074 rows equally distributed between 2 classes - suicide and non-suicide.
|
| 6 |
+
|
| 7 |
+
## Parameters
|
| 8 |
+
The model fine-tuning was conducted on 1 epoch, with batch size of 6, and learning rate of 0.00001. Due to limited computing resources and time, we were unable to scale up the number of epochs and batch size.
|
| 9 |
+
|
| 10 |
+
## Performance
|
| 11 |
+
The model has achieved the following results after fine-tuning on the aforementioned dataset:
|
| 12 |
+
- Accuracy: 0.9792
|
| 13 |
+
- Recall: 0.9788
|
| 14 |
+
- Precision: 0.9677
|
| 15 |
+
- F1 Score: 0.9732
|
| 16 |
+
|
| 17 |
+
## How to Use
|
| 18 |
+
Load the model via the transformers library:
|
| 19 |
+
```
|
| 20 |
+
from transformers import AutoTokenizer, AutoModel
|
| 21 |
+
tokenizer = AutoTokenizer.from_pretrained("gooohjy/suicidal-electra")
|
| 22 |
+
model = AutoModel.from_pretrained("gooohjy/suicidal-electra")
|
| 23 |
+
```
|
| 24 |
+
|
| 25 |
+
## Resources
|
| 26 |
+
For more resources, including the source code, please refer to the GitHub repository [gohjiayi/suicidal-text-detection](https://github.com/gohjiayi/suicidal-text-detection/).
|