Update README.md
Browse files
README.md
CHANGED
|
@@ -22,7 +22,6 @@ This model is a fine-tuned version of [distilbert-base-uncased](https://huggingf
|
|
| 22 |
The data is cleaned by selecting relevant columns and filtering rows based on whether they are labeled as 'accepted' or 'rejected'. It then groups the data by a unique identifier, concatenates text entries within each group into paragraphs, and prepares these paragraphs as predictors (X). Target labels (y) are derived from the final submission grade, mapping 'accepted' to 'violation' and 'rejected' to 'non-violation'. Finally, the data is split into training and testing sets using stratified sampling with a 20% test size and a random state of 1 for reproducibility.
|
| 23 |
|
| 24 |
It achieves the following results on the evaluation set:
|
| 25 |
-
- Loss: 1.0736
|
| 26 |
- Accuracy: 0.6937
|
| 27 |
- Precision: 0.6916
|
| 28 |
- Recall: 0.6937
|
|
|
|
| 22 |
The data is cleaned by selecting relevant columns and filtering rows based on whether they are labeled as 'accepted' or 'rejected'. It then groups the data by a unique identifier, concatenates text entries within each group into paragraphs, and prepares these paragraphs as predictors (X). Target labels (y) are derived from the final submission grade, mapping 'accepted' to 'violation' and 'rejected' to 'non-violation'. Finally, the data is split into training and testing sets using stratified sampling with a 20% test size and a random state of 1 for reproducibility.
|
| 23 |
|
| 24 |
It achieves the following results on the evaluation set:
|
|
|
|
| 25 |
- Accuracy: 0.6937
|
| 26 |
- Precision: 0.6916
|
| 27 |
- Recall: 0.6937
|