a

app.py

@@ -1,20 +1,30 @@
 import os
-import numpy as np
+#import numpy as np
 import gradio as gr
 import whisper
 import requests 
+import tempfile
+from neon_tts_plugin_coqui import CoquiTTS
 
+# Whisper: Speech-to-text
 model = whisper.load_model("base")
-##Bloom
+
+# The LLM : Bloom
 API_URL = "https://api-inference.huggingface.co/models/bigscience/bloom"
 HF_TOKEN = os.environ["HF_TOKEN"]
 headers = {"Authorization": f"Bearer {HF_TOKEN}"}
 
+# Text-to-Speech
+LANGUAGES = list(CoquiTTS.langs.keys())
+coquiTTS = CoquiTTS()
+
 
-def fun(audio) : #, state=''):
+# Processing input Audio
+def fun(audio) : 
   text1 = model.transcribe(audio)["text"]
   text2 = lang_model_response(text1)
-  return text1, text2
+  speech = tts(text, language):
+  return text1, text2, speech
   
 def lang_model_response(prompt): 
   print(f"*****Inside meme_generate - Prompt is :{prompt}")
@@ -40,17 +50,23 @@ def lang_model_response(prompt):
   print(f"output is : {output}") 
   output_tmp = output[0]['generated_text']
   print(f"output_tmp is: {output_tmp}")
-  solution = output_tmp[0]   #output_tmp.split("\nQ:")[0]   
+  solution = output_tmp.split(".")[1]   
   print(f"Final response after splits is: {solution}")
-    
-  #meme_image, new_prompt = write_on_image(solution)  
   return solution
-  
-def fun1(audio, state=''):
-  text = model.transcribe(audio)["text"]
-  state += text + " "
-  return state, state
-  
+
+#Text-to-Speech
+def tts(text, language):
+    with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as fp:
+        coquiTTS.get_tts(text, fp, speaker = {"language" : language})
+        return fp.name
+
+#inputs = [gr.Textbox(label="Input", value=CoquiTTS.langs["en"]["sentence"], max_lines=3), 
+#            gr.Radio(label="Language", choices=LANGUAGES, value="en")]
+#outputs = gr.Audio(label="Output")
+
+demo = gr.Interface(fn=tts, inputs=inputs, outputs=outputs)
+
+demo.launch()  
 gr.Interface(
     title = 'Testing Whisper', 
     fn=fun, 
@@ -59,6 +75,6 @@ gr.Interface(
        # "state"
     ],
     outputs=[
-        "textbox",  "textbox"
+        "textbox",  "textbox", "audio",
     ],
     live=True).launch()