Update run.sh

run.sh

@@ -1,25 +1,16 @@
 #!/bin/bash
 
-# Exit immediately if a command exits with a non-zero status.
-set -e
-
-# Start the Ollama server in the background.
-# The OLLAMA_HOST environment variable is already set in the Dockerfile, so it will listen on all interfaces.
+# Start the Ollama server in the background
 ollama serve &
 
-# Wait for the Ollama server to become available.
-echo "Waiting for Ollama to start..."
-while ! curl -s http://localhost:11434 > /dev/null; do
-  sleep 1
-done
-echo "Ollama started."
+# Wait a few seconds for the server to be fully operational
+sleep 5
 
-# Create the model from the Modelfile. This registers the downloaded GGUF file with Ollama.
-# 'gemma-unsloth' is the name we will use to refer to this model in our application.
-echo "Creating model 'create gemma3_4b_it_qat'..."
-ollama create gemma3_4b_it_qat -f Modelfile
-echo "Model created."
+# Pull the model from Hugging Face.
+# This command downloads the model and makes it available to the API.
+# It's the non-interactive version of "ollama run".
+ollama pull hf.co/unsloth/gemma-3-4b-it-qat-GGUF:Q4_K_M
 
-# Run the Gradio application.
-# It will connect to the Ollama server running on localhost.
-python app.py
+# Start the Gradio web application
+# This will connect to the Ollama server which is already running.
+python3 app.py