update speech2text module

app.py

@@ -79,33 +79,44 @@ def speech2text(audio):
 model_checkpoint = "huynguyen208/fantastic4-finetuned-vi-to-en-PhoMT-demo"
 translator = pipeline("translation", model=model_checkpoint)
 
-def translate(Vietnamese):
+def translate_vi2en(Vietnamese):
     return translator(Vietnamese)[0]['translation_text']
 
 def inference(audio):
     vi_text = speech2text(audio)
-    en_text = translate(vi_text)
+    en_text = translate_vi2en(vi_text)
     return en_text
 
 
 """Gradio demo"""
-iface1 = gr.Interface(fn=translate,
-                     inputs=["text"],
-                     outputs="text",
-                     title = 'Translate Vietnamese to English',
-                     description = 'Mini Translator')
-iface1.launch(inline = False)
 
+vi_example_text = ["Có phải bạn đang muốn tìm mua nhà ở ngoại ô thành phố Hồ Chí Minh không?",
+                   "có phải bạn đang muốn tìm mua nhà ở ngoại ô thành phố Hồ Chí Minh không"]
 
-inputs = gr.inputs.Audio(label="Input Audio", type="file")
-outputs =  gr.outputs.Textbox(label="Output Text")
-title = "Speech to text and translate Vietnamese to English"
-description = "Gradio demo for a wav2vec2-base-vietnamese-250h and Helsinki-NLP/opus-mt-vi-en"
-examples=[['vi_speech_01.wav'], ['vi_speech_02.wav'], ['vi_speech_03.wav']]
-iface2 = gr.Interface(inference,
-            inputs,
-            outputs, 
-            title=title,
-            description=description,
-            examples=examples)
-iface2.launch()
+vi_example_voice =[['vi_speech_01.wav'], ['vi_speech_02.wav'], ['vi_speech_03.wav']]
+
+with gr.Blocks() as demo:
+    with gr.Tabs():
+        with gr.TabItem("Translation: Vietnamese to English"):
+            with gr.Row():
+                with gr.Column():
+                    vietnamese = gr.Textbox(label="Vietnamese Text")
+                    translate_to_english = gr.Button(value="Translate To English")
+                with gr.Column():
+                    english = gr.Textbox(label="English Text")
+            translate_to_english.click(lambda text: translate_vi2en(text), inputs=vietnamese, outputs=english)
+            gr.Examples(examples=vi_example_text,
+                        inputs=[vietnamese])
+        with gr.TabItem("Speech2text translation: Vietnamese voice to English"):
+            with gr.Row():
+                inputs = gr.inputs.Audio(label="Input Audio", type="file")
+                outputs =  gr.outputs.Textbox(label="Output Text")
+                title = "Speech to text and translate Vietnamese to English"
+                description = "Gradio demo for a wav2vec2-base-vietnamese-250h and Helsinki-NLP/opus-mt-vi-en"
+                iface = gr.Interface(inference,
+                            inputs,
+                            outputs, 
+                            title=title,
+                            description=description,
+                            examples=vi_example_voice)
+                iface.launch()