Update README.md
Browse files
README.md
CHANGED
@@ -33,16 +33,15 @@ This model is intended for educational and demonstrational purposes only. It tak
|
|
33 |
- **Data Bias:** The model's knowledge is limited to the `gretelai/symptom_to_diagnosis` dataset. It cannot predict any condition outside of its 22-class training data and may perform poorly on symptom descriptions that are stylistically different from the training set.
|
34 |
- **Correlation, Not Causation:** The model learns statistical correlations between words and labels. It has no true understanding of biology or medicine.
|
35 |
|
36 |
-
## How to Use
|
37 |
-
|
38 |
-
To use this model, you must load the feature extractor (`Bio_ClinicalBERT`), the LightGBM classifier, and the label encoder.
|
39 |
-
|
40 |
## Training Data
|
41 |
This model was trained on the gretelai/symptom_to_diagnosis dataset, which contains ~1000 symptom descriptions across 22 balanced classes.
|
42 |
|
43 |
## Evaluation
|
44 |
The model achieves a Macro F1-score of 0.834 and an Accuracy of 0.835 on the test set.
|
45 |
|
|
|
|
|
|
|
46 |
```python
|
47 |
import torch
|
48 |
import joblib
|
|
|
33 |
- **Data Bias:** The model's knowledge is limited to the `gretelai/symptom_to_diagnosis` dataset. It cannot predict any condition outside of its 22-class training data and may perform poorly on symptom descriptions that are stylistically different from the training set.
|
34 |
- **Correlation, Not Causation:** The model learns statistical correlations between words and labels. It has no true understanding of biology or medicine.
|
35 |
|
|
|
|
|
|
|
|
|
36 |
## Training Data
|
37 |
This model was trained on the gretelai/symptom_to_diagnosis dataset, which contains ~1000 symptom descriptions across 22 balanced classes.
|
38 |
|
39 |
## Evaluation
|
40 |
The model achieves a Macro F1-score of 0.834 and an Accuracy of 0.835 on the test set.
|
41 |
|
42 |
+
## How to Use
|
43 |
+
To use this model, you must load the feature extractor (`Bio_ClinicalBERT`), the LightGBM classifier, and the label encoder.
|
44 |
+
|
45 |
```python
|
46 |
import torch
|
47 |
import joblib
|