Spaces:
Sleeping
Sleeping
customizing layout + adding html
Browse files- src/main.py +27 -5
- src/templates/index.html +10 -0
src/main.py
CHANGED
|
@@ -7,6 +7,7 @@ from transformers import AutoModel, AutoTokenizer
|
|
| 7 |
from .utils import extract_hidden_state
|
| 8 |
|
| 9 |
|
|
|
|
| 10 |
models_dir = os.path.join(os.path.dirname(__file__), '..', 'models')
|
| 11 |
model_file = os.path.join(models_dir, 'logistic_regression.pkl')
|
| 12 |
|
|
@@ -16,21 +17,42 @@ if os.path.exists(model_file):
|
|
| 16 |
else:
|
| 17 |
print(f"Error: {model_file} not found.")
|
| 18 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
model_name = "moussaKam/AraBART"
|
| 20 |
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
| 21 |
language_model = AutoModel.from_pretrained(model_name)
|
| 22 |
|
|
|
|
| 23 |
def classify_arabic_dialect(text):
|
| 24 |
text_embeddings = extract_hidden_state(text, tokenizer, language_model)
|
| 25 |
predicted_class = model.predict(text_embeddings)[0]
|
| 26 |
|
| 27 |
return predicted_class
|
| 28 |
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 34 |
|
| 35 |
|
| 36 |
if __name__ == "__main__":
|
|
|
|
| 7 |
from .utils import extract_hidden_state
|
| 8 |
|
| 9 |
|
| 10 |
+
# Load model
|
| 11 |
models_dir = os.path.join(os.path.dirname(__file__), '..', 'models')
|
| 12 |
model_file = os.path.join(models_dir, 'logistic_regression.pkl')
|
| 13 |
|
|
|
|
| 17 |
else:
|
| 18 |
print(f"Error: {model_file} not found.")
|
| 19 |
|
| 20 |
+
# Load html
|
| 21 |
+
html_dir = os.path.join(os.path.dirname(__file__), "templates")
|
| 22 |
+
index_html_path = os.path.join(html_dir, "index.html")
|
| 23 |
+
|
| 24 |
+
if os.path.exists(index_html_path):
|
| 25 |
+
with open(index_html_path, "r") as html_file:
|
| 26 |
+
index_html = html_file.read()
|
| 27 |
+
else:
|
| 28 |
+
print(f"Error: {index_html_path} not found.")
|
| 29 |
+
|
| 30 |
+
# Load pre-trained model
|
| 31 |
model_name = "moussaKam/AraBART"
|
| 32 |
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
| 33 |
language_model = AutoModel.from_pretrained(model_name)
|
| 34 |
|
| 35 |
+
|
| 36 |
def classify_arabic_dialect(text):
|
| 37 |
text_embeddings = extract_hidden_state(text, tokenizer, language_model)
|
| 38 |
predicted_class = model.predict(text_embeddings)[0]
|
| 39 |
|
| 40 |
return predicted_class
|
| 41 |
|
| 42 |
+
|
| 43 |
+
with gr.Blocks() as demo:
|
| 44 |
+
gr.HTML(index_html)
|
| 45 |
+
input_text = gr.Textbox(label="Your Arabic Text")
|
| 46 |
+
submit_btn = gr.Button("Submit")
|
| 47 |
+
with gr.Row():
|
| 48 |
+
first_country = gr.Textbox()
|
| 49 |
+
second_country = gr.Textbox()
|
| 50 |
+
third_country = gr.Textbox()
|
| 51 |
+
submit_btn.click(
|
| 52 |
+
fn=classify_arabic_dialect,
|
| 53 |
+
inputs=input_text,
|
| 54 |
+
outputs=[first_country, second_country, third_country])
|
| 55 |
+
gr.HTML("<p>Checkout the Github Repo:</p>")
|
| 56 |
|
| 57 |
|
| 58 |
if __name__ == "__main__":
|
src/templates/index.html
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
<!DOCTYPE html>
|
| 2 |
+
<html lang="en">
|
| 3 |
+
<head>
|
| 4 |
+
<title>Arabic Dialect Classifier</title>
|
| 5 |
+
</head>
|
| 6 |
+
<body>
|
| 7 |
+
<h1>Arabic Dialect Classifier</h1>
|
| 8 |
+
<p>Write some arabic text and get which dialect it is from</p>
|
| 9 |
+
</body>
|
| 10 |
+
</html>
|