Update app.py

app.py

@@ -1,9 +1,12 @@
+import os
+import requests
 import gradio as gr
-from huggingface_hub import InferenceClient
 from gtts import gTTS
 import tempfile
 
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
+
+TOGETHER_API_KEY = os.environ.get("TOGETHER_API_KEY")
+TOGETHER_MODEL = "meta-llama/Llama-3-70B-Instruct"
 
 def speak_text(text):
     tts = gTTS(text)
@@ -11,6 +14,27 @@ def speak_text(text):
         tts.save(fp.name)
         return fp.name
 
+def together_generate(prompt):
+    url = "https://api.together.xyz/v1/completions"
+    headers = {
+        "Authorization": f"Bearer {TOGETHER_API_KEY}",
+        "Content-Type": "application/json"
+    }
+    data = {
+        "model": TOGETHER_MODEL,
+        "prompt": prompt,
+        "max_tokens": 512,
+        "temperature": 0.7,
+        "top_p": 0.95,
+        "repetition_penalty": 1.0
+    }
+
+    response = requests.post(url, headers=headers, json=data)
+    if response.status_code == 200:
+        return response.json()['choices'][0]['text'].strip()
+    else:
+        return "⚠️ API error: " + response.text
+
 def respond(message, history):
     history = history or []
     prompt = ""
@@ -18,16 +42,7 @@ def respond(message, history):
         prompt += f"<|user|> {user_msg}\n<|assistant|> {bot_msg}\n"
     prompt += f"<|user|> {message}\n<|assistant|>"
 
-    response = client.text_generation(
-        prompt,
-        max_new_tokens=512,
-        temperature=0.7,
-        top_p=0.95,
-        repetition_penalty=1.0,
-        do_sample=True
-    )
-
-    response = response.strip().split("<|assistant|>")[-1].strip()
+    response = together_generate(prompt)
     history.append((message, response))
     audio_path = speak_text(response)
     return history, history, audio_path
@@ -35,16 +50,12 @@ def respond(message, history):
 with gr.Blocks() as demo:
     gr.Markdown("# Sambit AI 🤖")
     chatbot = gr.Chatbot([], label="Chat with AI", type="messages")
-    msg = gr.Textbox(label="Type your message here", placeholder="Ask me anything...", scale=7)
+    msg = gr.Textbox(label="Type your message here", placeholder="Ask me anything...")
     audio_output = gr.Audio(label="AI Voice", interactive=False)
-
-    clear = gr.Button("🧹 Clear Chat")
-
+    clear = gr.Button("Clear Chat")
     state = gr.State([])
 
     msg.submit(respond, [msg, state], [chatbot, state, audio_output])
     clear.click(lambda: ([], [], None), None, [chatbot, state, audio_output])
 
-# ✅ Required to make it work on Hugging Face
-demo.launch()
-
+demo.queue()