Spaces:
Sleeping
Sleeping
AreesaAshfaq
commited on
Commit
•
4438826
1
Parent(s):
ae981e1
Update app.py
Browse files
app.py
CHANGED
@@ -2,7 +2,7 @@ import streamlit as st
|
|
2 |
from transformers import MarianMTModel, MarianTokenizer
|
3 |
|
4 |
# Load the MarianMT model and tokenizer
|
5 |
-
model_name = "Helsinki-NLP/opus-mt-en-
|
6 |
model = MarianMTModel.from_pretrained(model_name)
|
7 |
tokenizer = MarianTokenizer.from_pretrained(model_name)
|
8 |
|
@@ -13,6 +13,14 @@ language_codes = {
|
|
13 |
'Italian': 'it',
|
14 |
'Portuguese': 'pt',
|
15 |
'Spanish': 'es',
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
16 |
# Add more languages if needed
|
17 |
}
|
18 |
|
@@ -20,12 +28,12 @@ def translate_text(text, target_lang_code):
|
|
20 |
# Prepare the input and translate
|
21 |
inputs = tokenizer.encode(text, return_tensors="pt")
|
22 |
# Generate the translation
|
23 |
-
translated = model.generate(inputs, decoder_start_token_id=tokenizer.
|
24 |
translated_text = tokenizer.decode(translated[0], skip_special_tokens=True)
|
25 |
return translated_text
|
26 |
|
27 |
def main():
|
28 |
-
st.title("
|
29 |
|
30 |
# User input for English text
|
31 |
text_to_translate = st.text_area("Enter text in English:")
|
|
|
2 |
from transformers import MarianMTModel, MarianTokenizer
|
3 |
|
4 |
# Load the MarianMT model and tokenizer
|
5 |
+
model_name = "Helsinki-NLP/opus-mt-en-many-to-many-mmt"
|
6 |
model = MarianMTModel.from_pretrained(model_name)
|
7 |
tokenizer = MarianTokenizer.from_pretrained(model_name)
|
8 |
|
|
|
13 |
'Italian': 'it',
|
14 |
'Portuguese': 'pt',
|
15 |
'Spanish': 'es',
|
16 |
+
'Chinese': 'zh',
|
17 |
+
'Japanese': 'ja',
|
18 |
+
'Korean': 'ko',
|
19 |
+
'Russian': 'ru',
|
20 |
+
'Arabic': 'ar',
|
21 |
+
'Turkish': 'tr',
|
22 |
+
'Hindi': 'hi',
|
23 |
+
'Swahili': 'sw',
|
24 |
# Add more languages if needed
|
25 |
}
|
26 |
|
|
|
28 |
# Prepare the input and translate
|
29 |
inputs = tokenizer.encode(text, return_tensors="pt")
|
30 |
# Generate the translation
|
31 |
+
translated = model.generate(inputs, decoder_start_token_id=tokenizer.lang_code_to_id[target_lang_code])
|
32 |
translated_text = tokenizer.decode(translated[0], skip_special_tokens=True)
|
33 |
return translated_text
|
34 |
|
35 |
def main():
|
36 |
+
st.title("Multilingual Translator")
|
37 |
|
38 |
# User input for English text
|
39 |
text_to_translate = st.text_area("Enter text in English:")
|