Create app.py

app.py

@@ -0,0 +1,50 @@
+import streamlit as st
+from llama_cpp import Llama
+from huggingface_hub import snapshot_download
+import os
+
+# Define the model repository and filename
+repo_id = "Divyansh12/check"
+filename = "unsloth.F16.gguf"
+
+# Download the model if not already present
+model_dir = "model"
+if not os.path.exists(model_dir):
+    st.write("Downloading the model...")
+    snapshot_download(repo_id=repo_id, cache_dir=model_dir)
+    st.write("Model downloaded successfully!")
+
+# Since the model file is directly inside the downloaded directory
+# Find the exact file path
+model_path = os.path.join(model_dir, filename)
+
+# Check if the file exists at the specified location
+if not os.path.exists(model_path):
+    st.error(f"Model file not found at {model_path}")
+else:
+    st.write(f"Found model file at {model_path}")
+
+    # Load the GGUF model using llama-cpp
+    st.write("Loading the model...")
+    llm = Llama.from_pretrained(
+        model_path=model_path
+    )
+    st.write("Model loaded successfully!")
+
+    # Streamlit input for the user to enter a prompt
+    prompt = st.text_input("Enter your prompt:", "What is the capital of France?")
+
+    # Generate the response when a prompt is given
+    if st.button("Generate Response"):
+        with st.spinner("Generating..."):
+            response = llm.create_chat_completion(
+                messages=[
+                    {
+                        "role": "user",
+                        "content": prompt
+                    }
+                ]
+            )
+            # Extract the message content from the response and display it
+            st.write("Response:")
+            st.write(response['choices'][0]['message']['content'])