Update app.py

app.py

@@ -1,35 +1,39 @@
 import streamlit as st
-from transformers import AutoModel, AutoTokenizer
-import torch
+from llama_cpp import Llama
 
 # Title for your app
-st.title("Llama-3-8B-Physics Master - Model Inference")
+st.title("Llama-3-8B-Physics Master - Chatbot")
 
-# Load the model and tokenizer from Hugging Face
+# Load the model from Hugging Face using llama_cpp
 @st.cache_resource
 def load_model():
-    model = AutoModel.from_pretrained("gallen881/Llama-3-8B-Physics_Master-GGUF")
-    tokenizer = AutoTokenizer.from_pretrained("gallen881/Llama-3-8B-Physics_Master-GGUF")
-    return model, tokenizer
+    # Load the model from the Hugging Face Hub
+    model = Llama.from_pretrained(
+        repo_id="gallen881/Llama-3-8B-Physics_Master-GGUF",
+        filename="unsloth.F16.gguf"  # or unsloth.Q4_K_M.gguf for a smaller file
+    )
+    return model
 
 # Load the model once and store it in cache
-model, tokenizer = load_model()
+model = load_model()
 
 # Text input for the user
-user_input = st.text_area("Enter your input here:")
+user_input = st.text_area("Enter your message here:")
 
-if st.button("Generate Output"):
+if st.button("Generate Response"):
     if user_input:
-        # Tokenize the input
-        inputs = tokenizer(user_input, return_tensors="pt")
-        
-        # Forward pass through the model
-        with torch.no_grad():
-            outputs = model(**inputs)
-        
-        # Get the output embeddings or logits (depending on the model structure)
-        # For example, let's say we want to display embeddings
-        st.write("Model Output Embeddings:", outputs.last_hidden_state)
+        # Create chat completion with the model
+        response = model.create_chat_completion(
+            messages=[
+                {
+                    "role": "user",
+                    "content": user_input
+                }
+            ]
+        )
+
+        # Extract the content from the model's response
+        st.write("Model Response:", response['choices'][0]['message']['content'])
 
     else:
-        st.write("Please enter some input.")
+        st.write("Please enter a message.")