Update app.py

app.py

@@ -1,13 +1,13 @@
 import os
 import re
 import gradio as gr
-import edge_tts
 import asyncio
 import time
 import tempfile
 from huggingface_hub import InferenceClient
+from gtts import gTTS
 
-DESCRIPTION = """ # <center><b>JARVIS⚡</b></center>
+DESCRIPTION = """ # <center><b>JARVIS⚡ 수정본 </b></center>
         ### <center>A personal Assistant of Tony Stark for YOU
         ### <center>Currently It supports text input, But If this space completes 1k hearts than I starts working on Audio Input.</center>
         """
@@ -43,11 +43,11 @@ async def generate1(prompt):
     for response in stream:
         output += response.token.text
 
-    communicate = edge_tts.Communicate(output)
-    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp_file:
-        tmp_path = tmp_file.name
-        await communicate.save(tmp_path)
-    yield tmp_path
+  tts = gTTS(text=output, lang='ko')
+with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp_file:
+tmp_path = tmp_file.name
+tts.save(tmp_path)
+yield tmp_path
 
 client2 = InferenceClient("meta-llama/Meta-Llama-3-70B-Instruct")
 
@@ -62,17 +62,17 @@ async def generate2(prompt):
         do_sample=True,
     )    
     formatted_prompt = system_instructions2 + prompt + "[ASSISTANT]"
-    stream = client2.text_generation(
+    stream = client3.text_generation(
         formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=True)
     output = ""
     for response in stream:
         output += response.token.text
 
-    communicate = edge_tts.Communicate(output)
-    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp_file:
-        tmp_path = tmp_file.name
-        await communicate.save(tmp_path)
-    yield tmp_path
+    tts = gTTS(text=output, lang='ko')
+with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp_file:
+tmp_path = tmp_file.name
+tts.save(tmp_path)
+yield tmp_path
 
 client3 = InferenceClient("meta-llama/Meta-Llama-3-70B-Instruct")
 
@@ -93,11 +93,11 @@ async def generate3(prompt):
     for response in stream:
         output += response.token.text
 
-    communicate = edge_tts.Communicate(output)
-    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp_file:
-        tmp_path = tmp_file.name
-        await communicate.save(tmp_path)
-    yield tmp_path
+tts = gTTS(text=output, lang='ko')
+with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp_file:
+tmp_path = tmp_file.name
+tts.save(tmp_path)
+yield tmp_path
 
 with gr.Blocks(css="style.css") as demo:    
     gr.Markdown(DESCRIPTION)