Update README.md
Browse files
README.md
CHANGED
@@ -31,6 +31,23 @@ This set contains two models trained in the context of a thesis on emotion recog
|
|
31 |
|
32 |
Both models are in `.h5` format and were developed using TensorFlow and Keras. They have been fine-tuned to capture emotional expressions relevant to educational environments.
|
33 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
34 |
This resource has been shared for scientific reproducibility and to support academic and technical analysis.
|
35 |
|
36 |
## Source Code Repository
|
|
|
31 |
|
32 |
Both models are in `.h5` format and were developed using TensorFlow and Keras. They have been fine-tuned to capture emotional expressions relevant to educational environments.
|
33 |
|
34 |
+
|
35 |
+
**Note on Model Loading**:
|
36 |
+
- The **FacialModel** can be loaded using standard Keras methods:
|
37 |
+
```python
|
38 |
+
model = tf.keras.models.load_model("FacialModel-8177.h5")
|
39 |
+
```
|
40 |
+
- The **TextModel** uses custom objects (e.g., `TFRobertaModel`). You must specify them when loading the model:
|
41 |
+
```python
|
42 |
+
from transformers import TFRobertaModel
|
43 |
+
from tensorflow.keras.models import load_model
|
44 |
+
|
45 |
+
def custom_objects_fn_roberta():
|
46 |
+
return {'TFRobertaModel': TFRobertaModel}
|
47 |
+
|
48 |
+
model = load_model("TextModel-0.7836.h5", custom_objects=custom_objects_fn_roberta())
|
49 |
+
```
|
50 |
+
|
51 |
This resource has been shared for scientific reproducibility and to support academic and technical analysis.
|
52 |
|
53 |
## Source Code Repository
|