Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -1,49 +1,49 @@
|
|
1 |
import streamlit as st
|
2 |
from transformers import MarianMTModel, MarianTokenizer
|
3 |
|
4 |
-
#
|
5 |
LANGUAGES = {
|
6 |
-
"English"
|
7 |
-
"French"
|
8 |
-
"German"
|
9 |
-
"Spanish"
|
10 |
-
"Italian"
|
11 |
-
"Portuguese"
|
12 |
-
"Russian"
|
13 |
-
"Chinese"
|
14 |
-
"
|
15 |
-
"Arabic"
|
16 |
}
|
17 |
|
18 |
# Helper function to load the model and tokenizer
|
19 |
@st.cache_resource
|
20 |
-
def load_translation_model(src_lang, tgt_lang):
|
21 |
-
model_name = f"Helsinki-NLP
|
22 |
tokenizer = MarianTokenizer.from_pretrained(model_name)
|
23 |
model = MarianMTModel.from_pretrained(model_name)
|
24 |
-
return tokenizer,model
|
25 |
|
26 |
# Translation function
|
27 |
def translate_text(tokenizer, model, text):
|
28 |
inputs = tokenizer(text, return_tensors='pt', padding=True)
|
29 |
translated = model.generate(**inputs)
|
30 |
-
return tokenizer.decode(translated[0],skip_special_tokens
|
31 |
|
32 |
-
#Streamlit App
|
33 |
st.title("Language Translation APP 🌍")
|
34 |
-
st.write("Translate text between multiple
|
35 |
|
36 |
-
#Language selection
|
37 |
col1, col2 = st.columns(2)
|
38 |
with col1:
|
39 |
source_language = st.selectbox("Select Source Language 🌐", list(LANGUAGES.keys()))
|
40 |
with col2:
|
41 |
-
target_language = st.selectbox("Select Target Language 🌍",list(LANGUAGES.keys()))
|
42 |
|
43 |
-
#Input Text
|
44 |
-
text_to_translate
|
45 |
|
46 |
-
#Translate button
|
47 |
if st.button("Translate"):
|
48 |
if source_language == target_language:
|
49 |
st.warning("Source and target language must be different")
|
@@ -54,16 +54,15 @@ if st.button("Translate"):
|
|
54 |
tgt_lang = LANGUAGES[target_language]
|
55 |
|
56 |
try:
|
57 |
-
#Load model and tokenizer
|
58 |
tokenizer, model = load_translation_model(src_lang, tgt_lang)
|
59 |
|
60 |
-
#Perform translation
|
61 |
translated_text = translate_text(tokenizer, model, text_to_translate)
|
62 |
|
63 |
# Display result
|
64 |
-
st.subheader("
|
65 |
-
st.write(
|
66 |
|
67 |
except Exception as e:
|
68 |
-
st.error(f"Error
|
69 |
-
|
|
|
1 |
import streamlit as st
|
2 |
from transformers import MarianMTModel, MarianTokenizer
|
3 |
|
4 |
+
# Specified different languages
|
5 |
LANGUAGES = {
|
6 |
+
"English": "en",
|
7 |
+
"French": "fr",
|
8 |
+
"German": "de",
|
9 |
+
"Spanish": "es",
|
10 |
+
"Italian": "it",
|
11 |
+
"Portuguese": "pt",
|
12 |
+
"Russian": "ru",
|
13 |
+
"Chinese": "zh",
|
14 |
+
"Japanese": "ja",
|
15 |
+
"Arabic": "ar",
|
16 |
}
|
17 |
|
18 |
# Helper function to load the model and tokenizer
|
19 |
@st.cache_resource
|
20 |
+
def load_translation_model(src_lang, tgt_lang): # Corrected parameter names
|
21 |
+
model_name = f"Helsinki-NLP/opus-mt-{src_lang}-{tgt_lang}" # Corrected typo in model_name
|
22 |
tokenizer = MarianTokenizer.from_pretrained(model_name)
|
23 |
model = MarianMTModel.from_pretrained(model_name)
|
24 |
+
return tokenizer, model
|
25 |
|
26 |
# Translation function
|
27 |
def translate_text(tokenizer, model, text):
|
28 |
inputs = tokenizer(text, return_tensors='pt', padding=True)
|
29 |
translated = model.generate(**inputs)
|
30 |
+
return tokenizer.decode(translated[0], skip_special_tokens=True)
|
31 |
|
32 |
+
# Streamlit App
|
33 |
st.title("Language Translation APP 🌍")
|
34 |
+
st.write("Translate text between multiple languages using open-source models.")
|
35 |
|
36 |
+
# Language selection
|
37 |
col1, col2 = st.columns(2)
|
38 |
with col1:
|
39 |
source_language = st.selectbox("Select Source Language 🌐", list(LANGUAGES.keys()))
|
40 |
with col2:
|
41 |
+
target_language = st.selectbox("Select Target Language 🌍", list(LANGUAGES.keys()))
|
42 |
|
43 |
+
# Input Text
|
44 |
+
text_to_translate = st.text_area("Enter text to translate", height=150)
|
45 |
|
46 |
+
# Translate button
|
47 |
if st.button("Translate"):
|
48 |
if source_language == target_language:
|
49 |
st.warning("Source and target language must be different")
|
|
|
54 |
tgt_lang = LANGUAGES[target_language]
|
55 |
|
56 |
try:
|
57 |
+
# Load model and tokenizer
|
58 |
tokenizer, model = load_translation_model(src_lang, tgt_lang)
|
59 |
|
60 |
+
# Perform translation
|
61 |
translated_text = translate_text(tokenizer, model, text_to_translate)
|
62 |
|
63 |
# Display result
|
64 |
+
st.subheader("Translated Text 🔄:")
|
65 |
+
st.write(translated_text)
|
66 |
|
67 |
except Exception as e:
|
68 |
+
st.error(f"Error: {str(e)}")
|
|