Back to InferenceClient with Mixtral-8x7B

app.py

@@ -1,35 +1,36 @@
 import os
-import requests
 import streamlit as st
+from huggingface_hub import InferenceClient
 
 st.set_page_config(page_title="Recipe Generator", layout="centered")
 
 st.title("🍳 AI Recipe Generator")
 st.write("Enter ingredients and get a recipe suggestion!")
 
-API_URL = "https://api-inference.huggingface.co/models/HuggingFaceH4/zephyr-7b-beta/v1/chat/completions"
+@st.cache_resource
+def get_client():
+    token = os.environ.get("HF_TOKEN")
+    return InferenceClient(token=token)
+
+client = get_client()
 
 ingredients = st.text_input("Enter ingredients (comma-separated):", placeholder="eggs, cheese, spinach")
 
 if st.button("Generate Recipe"):
     if ingredients:
         with st.spinner("Generating recipe..."):
-            token = os.environ.get("HF_TOKEN")
-            headers = {"Authorization": f"Bearer {token}"}
             prompt = (
                 f"Write a complete recipe using these ingredients: {ingredients}. "
                 "Include: a creative recipe name, ingredients with amounts, "
                 "numbered cooking steps, and estimated cooking time."
             )
-            payload = {
-                "model": "HuggingFaceH4/zephyr-7b-beta",
-                "messages": [{"role": "user", "content": prompt}],
-                "max_tokens": 500,
-                "temperature": 0.7,
-            }
-            response = requests.post(API_URL, headers=headers, json=payload)
-            response.raise_for_status()
-            recipe = response.json()["choices"][0]["message"]["content"]
+            response = client.chat_completion(
+                model="mistralai/Mixtral-8x7B-Instruct-v0.1",
+                messages=[{"role": "user", "content": prompt}],
+                max_tokens=500,
+                temperature=0.7,
+            )
+            recipe = response.choices[0].message.content
             st.success(recipe)
     else:
         st.warning("Please enter some ingredients!")

requirements.txt

@@ -1,2 +1,2 @@
 streamlit>=1.28.0
-requests>=2.31.0
+huggingface_hub>=0.23.0