Spaces:
Build error
Build error
frogcho123
commited on
Commit
•
e6cfad1
1
Parent(s):
97ed968
Update app.py
Browse files
app.py
CHANGED
@@ -1,5 +1,4 @@
|
|
1 |
import gradio as gr
|
2 |
-
import os
|
3 |
import whisper
|
4 |
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
|
5 |
from gtts import gTTS
|
@@ -14,7 +13,7 @@ def translate_audio(input_file, target_language):
|
|
14 |
temp_audio_path = temp_audio.name
|
15 |
|
16 |
# Auto to text (STT)
|
17 |
-
model = whisper.
|
18 |
audio = whisper.load_audio(temp_audio_path)
|
19 |
audio = whisper.pad_or_trim(audio)
|
20 |
mel = whisper.log_mel_spectrogram(audio).to(model.device)
|
@@ -27,7 +26,8 @@ def translate_audio(input_file, target_language):
|
|
27 |
# Translate
|
28 |
tokenizer = AutoTokenizer.from_pretrained("alirezamsh/small100")
|
29 |
model = AutoModelForSeq2SeqLM.from_pretrained("alirezamsh/small100")
|
30 |
-
tokenizer.src_lang =
|
|
|
31 |
encoded_bg = tokenizer(text, return_tensors="pt")
|
32 |
generated_tokens = model.generate(**encoded_bg)
|
33 |
translated_text = tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)[0]
|
@@ -41,12 +41,12 @@ def translate_audio(input_file, target_language):
|
|
41 |
|
42 |
# Define Gradio interface
|
43 |
inputs = [
|
44 |
-
|
45 |
-
|
46 |
]
|
47 |
|
48 |
outputs = [
|
49 |
-
|
50 |
]
|
51 |
|
52 |
title = "Audio Translation"
|
|
|
1 |
import gradio as gr
|
|
|
2 |
import whisper
|
3 |
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
|
4 |
from gtts import gTTS
|
|
|
13 |
temp_audio_path = temp_audio.name
|
14 |
|
15 |
# Auto to text (STT)
|
16 |
+
model = whisper.Whisper("base")
|
17 |
audio = whisper.load_audio(temp_audio_path)
|
18 |
audio = whisper.pad_or_trim(audio)
|
19 |
mel = whisper.log_mel_spectrogram(audio).to(model.device)
|
|
|
26 |
# Translate
|
27 |
tokenizer = AutoTokenizer.from_pretrained("alirezamsh/small100")
|
28 |
model = AutoModelForSeq2SeqLM.from_pretrained("alirezamsh/small100")
|
29 |
+
tokenizer.src_lang = lang
|
30 |
+
tokenizer.tgt_lang = target_language
|
31 |
encoded_bg = tokenizer(text, return_tensors="pt")
|
32 |
generated_tokens = model.generate(**encoded_bg)
|
33 |
translated_text = tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)[0]
|
|
|
41 |
|
42 |
# Define Gradio interface
|
43 |
inputs = [
|
44 |
+
gr.inputs.File(label="Upload Audio File"),
|
45 |
+
gr.inputs.Dropdown(choices=['en', 'es', 'fr', 'de', 'ru'], label="Target Language")
|
46 |
]
|
47 |
|
48 |
outputs = [
|
49 |
+
gr.outputs.File(label="Translated Audio")
|
50 |
]
|
51 |
|
52 |
title = "Audio Translation"
|