Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
from transformers import MBartForConditionalGeneration, MBart50TokenizerFast
|
3 |
+
from transformers import AutoModelForMaskedLM, AutoTokenizer
|
4 |
+
|
5 |
+
# Load the models and tokenizers
|
6 |
+
model_translation = MBartForConditionalGeneration.from_pretrained("facebook/mbart-large-50-one-to-many-mmt")
|
7 |
+
model_masked_lm = AutoModelForMaskedLM.from_pretrained("alabnii/jmedroberta-base-sentencepiece")
|
8 |
+
model_translation.eval()
|
9 |
+
|
10 |
+
tokenizer_translation = MBart50TokenizerFast.from_pretrained("facebook/mbart-large-50-one-to-many-mmt", src_lang="en_XX")
|
11 |
+
tokenizer_masked_lm = AutoTokenizer.from_pretrained("alabnii/jmedroberta-base-sentencepiece")
|
12 |
+
|
13 |
+
text = st.text_area('Enter the text:')
|
14 |
+
|
15 |
+
if text:
|
16 |
+
model_inputs = tokenizer_translation(text, return_tensors="pt")
|
17 |
+
lg = st.text_input("Select Language: 1.Hindi, 2.Telugu, 3.Gujarati, 4.Bengali")
|
18 |
+
if lg==1:
|
19 |
+
generated_tokens = model_translation.generate(
|
20 |
+
**model_inputs,
|
21 |
+
forced_bos_token_id=tokenizer_translation.lang_code_to_id["hi_IN"]
|
22 |
+
)
|
23 |
+
else if lg==2:
|
24 |
+
generated_tokens = model_translation.generate(
|
25 |
+
**model_inputs,
|
26 |
+
forced_bos_token_id=tokenizer_translation.lang_code_to_id["te_IN"]
|
27 |
+
)
|
28 |
+
else if lg==3:
|
29 |
+
generated_tokens = model_translation.generate(
|
30 |
+
**model_inputs,
|
31 |
+
forced_bos_token_id=tokenizer_translation.lang_code_to_id["gu_IN"]
|
32 |
+
)
|
33 |
+
else if lg==4:
|
34 |
+
generated_tokens = model_translation.generate(
|
35 |
+
**model_inputs,
|
36 |
+
forced_bos_token_id=tokenizer_translation.lang_code_to_id["bn_IN"]
|
37 |
+
)
|
38 |
+
else:
|
39 |
+
return 0
|
40 |
+
|
41 |
+
translation = tokenizer_translation.batch_decode(generated_tokens, skip_special_tokens=True)
|
42 |
+
st.json(translation)
|