Spaces:
Runtime error
Runtime error
Simon Salmon
commited on
Commit
·
67647ba
1
Parent(s):
78e9cc0
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
from transformers import MBartForConditionalGeneration, MBart50TokenizerFast
|
3 |
+
|
4 |
+
|
5 |
+
model_name = "ramsrigouthamg/t5_paraphraser"
|
6 |
+
model = T5ForConditionalGeneration.from_pretrained('ramsrigouthamg/t5_paraphraser')
|
7 |
+
tokenizer = T5Tokenizer.from_pretrained('ramsrigouthamg/t5_paraphraser')
|
8 |
+
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
9 |
+
model = model.to(device)
|
10 |
+
|
11 |
+
def translate_to_english(model, tokenizer, text):
|
12 |
+
translated_text = []
|
13 |
+
text = "paraphrase: " + text + " </s>"
|
14 |
+
encoding = tokenizer.encode_plus(text,pad_to_max_length=True, return_tensors="pt")
|
15 |
+
input_ids, attention_masks = encoding["input_ids"].to(device), encoding["attention_mask"].to(device)
|
16 |
+
beam_outputs = model.generate(
|
17 |
+
input_ids=input_ids, attention_mask=attention_masks,
|
18 |
+
do_sample=True,
|
19 |
+
max_length=256,
|
20 |
+
top_k=120,
|
21 |
+
top_p=0.98,
|
22 |
+
early_stopping=True,
|
23 |
+
num_return_sequences=10
|
24 |
+
)
|
25 |
+
for beam_output in beam_outputs:
|
26 |
+
sent = tokenizer.decode(beam_output, skip_special_tokens=True,clean_up_tokenization_spaces=True)
|
27 |
+
print(sent)
|
28 |
+
translated_text.append(sent)
|
29 |
+
return translated_text
|
30 |
+
|
31 |
+
st.title("Auto Translate (To English)")
|
32 |
+
text = st.text_input("Okay")
|
33 |
+
st.text("What you wrote: ")
|
34 |
+
st.write(text)
|
35 |
+
st.text("English Translation: ")
|
36 |
+
if text:
|
37 |
+
translated_text = translate_to_english(model, tokenizer, text)
|
38 |
+
st.write(translated_text if translated_text else "No translation found")
|