Update app.py

app.py

@@ -1,16 +1,13 @@
 import streamlit as st
-from transformers import AutoModelForCausalLM, AutoTokenizer
-import torch, re, json,huggingface_hub,os
+import re
+import json,os
+from openai import OpenAI
 from datetime import datetime
-huggingface_hub.login(token=os.getenv("huggingface_id"))
 
-# Load model and tokenizer
-model = AutoModelForCausalLM.from_pretrained(
-    "meta-llama/Meta-Llama-3.1-8B-Instruct",
-    torch_dtype="auto",
-    device_map="auto",
+client = OpenAI(
+    api_key = os.getenv("OPENAI-API-KEY")
 )
-tokenizer = AutoTokenizer.from_pretrained("meta-llama/Meta-Llama-3.1-8B-Instruct")
+
 
 if 'recipe' not in st.session_state:
     st.session_state.recipe = None
@@ -111,27 +108,14 @@ def generate_recipe(user_inputs):
         ]
         prompt = create_detailed_prompt(user_inputs['user_direction'], user_inputs['exclusions'], user_inputs['serving_size'], user_inputs['difficulty'])
         messages = [{"role": "user", "content": prompt}]
-        text = tokenizer.apply_chat_template(
-            messages,
-            tokenize=False,
-            add_generation_prompt=True,
-            tools=functions
+        st.session_state.recipe = client.chat.completions.create(
+          model="gpt-4-1106-preview",
+          messages=messages,
+          temperature=0.75,
+          top_p=0.75,
+          functions=functions,
+          function_call={"name":"provide_recipe"},  # auto is default, but we'll be explicit
         )
-
-        # Tokenize and move to the correct device
-        model_inputs = tokenizer([text], return_tensors="pt")
-        torch.cuda.empty_cache()
-        with torch.no_grad():
-            generated_ids = model.generate(
-                model_inputs.input_ids,
-                max_new_tokens=1024  # Reduced value for memory management
-            )
-
-        generated_ids = [
-            output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
-        ]
-
-        st.session_state.recipe = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
         st.session_state.recipe_saved = False
         
 def clear_inputs():