Update app.py

app.py

@@ -1,4 +1,4 @@
-import gradio as gr
+import streamlit as st
 import edge_tts
 import asyncio
 import tempfile
@@ -11,7 +11,7 @@ import random
 
 default_lang = "en"
 
-engines = { default_lang: Model(default_lang) }
+engines = {default_lang: Model(default_lang)}
 
 def transcribe(audio):
     lang = "en"
@@ -30,7 +30,7 @@ def client_fn(model):
         return InferenceClient("mistralai/Mistral-7B-Instruct-v0.3")
     elif "Phi" in model:
         return InferenceClient("microsoft/Phi-3-mini-4k-instruct")
-    else: 
+    else:
         return InferenceClient("microsoft/Phi-3-mini-4k-instruct")
 
 def randomize_seed_fn(seed: int) -> int:
@@ -48,15 +48,15 @@ Respond in a normal, conversational manner while being friendly and helpful.
 def models(text, model="Mixtral 8x7B", seed=42):
 
     seed = int(randomize_seed_fn(seed))
-    generator = torch.Generator().manual_seed(seed)  
-    
+    generator = torch.Generator().manual_seed(seed)
+
     client = client_fn(model)
-    
+
     generate_kwargs = dict(
         max_new_tokens=300,
         seed=seed
     )
-    
+
     formatted_prompt = system_instructions1 + text + "[JARVIS]"
     stream = client.text_generation(
         formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
@@ -74,69 +74,34 @@ async def respond(audio, model, seed):
     with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp_file:
         tmp_path = tmp_file.name
         await communicate.save(tmp_path)
-    yield tmp_path
-
-DESCRIPTION = """ # <center><b>JARVIS⚡</b></center>
-        ### <center>A personal Assistant of Tony Stark for YOU
-        ### <center>Voice Chat with your personal Assistant</center>
-        """
-
-with gr.Blocks(css="style.css") as demo:    
-    gr.Markdown(DESCRIPTION)
-    with gr.Row():
-        select = gr.Dropdown([ 'Mixtral 8x7B',
-        'Llama 3 8B',
-        'Mistral 7B v0.3',
-        'Phi 3 mini',
-    ],
-    value="Mistral 7B v0.3",
-    label="Model"
-    )
-        seed = gr.Slider(
-        label="Seed",
-        minimum=0,
-        maximum=999999,
-        step=1,
-        value=0,
-        visible=False
-        )
-        input = gr.Audio(label="User", sources="microphone", type="filepath", waveform_options=False)
-        output = gr.Audio(label="AI", type="filepath",
-                        interactive=False,
-                        autoplay=True,
-                        elem_classes="audio")
-        gr.Interface(
-            batch=True,
-            max_batch_size=10, 
-            fn=respond, 
-            inputs=[input, select, seed],
-            outputs=[output], live=True)  
-        
-    with gr.Row():
-        select = gr.Dropdown([ 'Mixtral 8x7B',
-        'Llama 3 8B',
-        'Mistral 7B v0.3',
-        'Phi 3 mini',
-    ],
-    value="Mistral 7B v0.3",
-    label="Model"
-    )
-        seed = gr.Slider(
-        label="Seed",
-        minimum=0,
-        maximum=999999,
-        step=1,
-        value=0,
-        visible=False
-        )
-        input = gr.Textbox(label="User")
-        output = gr.Textbox(label="AI", interactive=False)
-        gr.Interface(
-            batch=True,
-            max_batch_size=10, 
-            fn=models, 
-            inputs=[input, select, seed],
-            outputs=[output], live=True)  
-
-if __name__ == "__main__":
-    demo.queue(max_size=200).launch()
+    return tmp_path
+
+st.title("JARVIS⚡")
+st.markdown("### A personal Assistant of Tony Stark for YOU")
+st.markdown("### Voice Chat with your personal Assistant")
+
+with st.form("voice_form"):
+    model_choice = st.selectbox("Choose a model", ['Mixtral 8x7B', 'Llama 3 8B', 'Mistral 7B v0.3', 'Phi 3 mini'], key="voice_model")
+    audio_file = st.file_uploader("Upload Audio", type=["wav", "mp3"], key="voice_audio")
+    submit_button = st.form_submit_button("Submit")
+
+    if submit_button:
+        if audio_file is not None:
+            with st.spinner("Transcribing and generating response..."):
+                audio_bytes = audio_file.read()
+                with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp_file:
+                    tmp_file.write(audio_bytes)
+                    tmp_path = tmp_file.name
+                    response = respond(tmp_path, model_choice, 42)
+                    st.audio(response, format='audio/wav')
+
+with st.form("text_form"):
+    model_choice = st.selectbox("Choose a model", ['Mixtral 8x7B', 'Llama 3 8B', 'Mistral 7B v0.3', 'Phi 3 mini'], key="text_model")
+    user_text = st.text_area("Enter your message:", key="text_input")
+    submit_button = st.form_submit_button("Submit")
+
+    if submit_button:
+        if user_text:
+            with st.spinner("Generating response..."):
+                response = models(user_text, model_choice, 42)
+                st.text_area("JARVIS Response", value=response, key="text_output", height=150)