Upload 2 files

app.py

@@ -0,0 +1,35 @@
+import gradio as gr
+from googletrans import Translator
+from transformers import pipeline
+
+pipe = pipeline(model="Nikolajvestergaard/Japanese_Fine_Tuned_Whisper_Model")
+
+def translate_and_transcribe(audio):
+    translator = Translator()
+    
+    # Transcribe Japanese audio to text
+    transcription = pipe(audio)["text"]  
+
+    # Translate the transcription to English
+    result = translator.translate(transcription, dest='en')
+    translation = result.text
+
+    # Get the pronunciation of the transcription in Japanese
+    pronunciation = translator.translate(transcription, dest='ja').pronunciation
+
+    return transcription, pronunciation, translation
+
+input_audio = gr.inputs.Audio(label="Upload your Japanese speech here. Try to say 'Kon'nichiwa', 'Arigatō' or perhaps 'Sayōnara'", source="microphone", type="filepath")
+output_textbox1 = gr.outputs.Textbox(label="Transcription")
+output_textbox2 = gr.outputs.Textbox(label="Pronunciation")
+output_textbox3 = gr.outputs.Textbox(label="Translation")
+
+iface = gr.Interface(
+    fn=translate_and_transcribe, 
+    inputs=input_audio, 
+    outputs=[output_textbox1, output_textbox2, output_textbox3],
+    title="Japanese Automatic Speech Recognition, Pronunciation and Translation",
+    description="Record Japanese speech to get its pronunciation and translate it to English. All done by using a fine-tuned version of the tiny Whisper model which is connected to a Google Translate API"
+)
+
+iface.launch()

requirements.txt

@@ -0,0 +1,2 @@
+googletrans==4.0.0-rc1
+gradio