start.sh

#!/bin/bash

export OLLAMA_MODELS=/root/.ollama/models
export OLLAMA_KEEP_ALIVE=-1

ollama serve &
OLLAMA_PID=$!

echo "Waiting for Ollama to start..."
sleep 10

echo "Pulling qwen3:1.7b..."
ollama pull qwen3:1.7b

echo "Pulling qwen2.5:1.5b..."
ollama pull qwen2.5:1.5b

echo "Pulling granite4:3b-h..."
ollama pull granite4:3b-h

echo "Pulling lfm2.5-thinking:1.2b..."
ollama pull lfm2.5-thinking:1.2b

# حمّل الكل في RAM فوراً
echo "Loading all models into RAM..."
ollama run qwen3:1.7b "" &
ollama run qwen2.5:1.5b "" &
ollama run granite4:3b-h "" &
ollama run lfm2.5-thinking:1.2b "" &

echo "✅ All models ready!"
ollama list

wait $OLLAMA_PID