AI Tips and Tricks with Ollama

#Quick Start Guide

#Installation & Setup

Get Ollama running on your system in minutes.

# Linux/macOS
curl -fsSL https://ollama.ai/install.sh | sh

# Windows — Download from ollama.ai and run installer

#First Model Download

Start with a lightweight, versatile model.

# Terminal
ollama pull llama2:7b

This downloads the 7B parameter Llama 2 model (approx. 3.8GB).

#Test Your Setup

Verify everything works correctly.

# Terminal
ollama run llama2:7b "Hello, world!"

#Smart Model Management

List Available Models

ollama list

Remove Unused Models

ollama rm model-name:tag

Update Models

ollama pull model-name:tag

#Popular Model Recommendations

#Performance Optimization

#GPU Acceleration

Ollama automatically detects and uses GPU acceleration when available. Ensure you have CUDA (NVIDIA) or Metal (Apple) support.

# Check GPU usage
nvidia-smi # For NVIDIA
activity monitor # For Apple Silicon

#Memory Management

Set memory limits to prevent system overload:

# Set in environment
export OLLAMA_MAX_MEMORY=8GB

#Concurrent Requests

Configure parallel request handling:

# Set max concurrent requests
export OLLAMA_MAX_QUEUE=10

#Model Loading

Keep frequently used models in memory:

ollama run llama2:7b --keep-alive 30m

#API Integration Tips

#REST API Examples

Basic Chat Completion

# cURL
curl http://localhost:11434/api/generate \
  -d '{
    "model": "llama2:7b",
    "prompt": "Why is the sky blue?",
    "stream": false
  }'

Python Integration

import requests
import json

response = requests.post('http://localhost:11434/api/generate',
    json={
        'model': 'llama2:7b',
        'prompt': 'Explain quantum computing',
        'stream': False
    })

result = response.json()
print(result['response'])

Streaming Responses

const response = await fetch('http://localhost:11434/api/generate', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({
    model: 'llama2:7b',
    prompt: 'Write a story about AI',
    stream: true
  })
});

const reader = response.body.getReader();
while (true) {
  const { done, value } = await reader.read();
  if (done) break;

  const chunk = new TextDecoder().decode(value);
  const data = JSON.parse(chunk);
  console.log(data.response);
}

#Advanced Techniques

#Custom Model Files

Create specialized models with custom parameters.

# Modelfile
FROM llama2:7b
PARAMETER temperature 0.8
PARAMETER top_p 0.9
SYSTEM "You are a helpful coding assistant."

# Build Model
ollama create my-coding-assistant -f ./Modelfile

#Prompt Engineering

Optimize prompts for better responses.

• Be specific and clear in instructions
• Use examples for complex tasks
• Break down complex requests
• Specify output format when needed

#Context Window Management

Efficiently handle long conversations.

# Limit context length
ollama run llama2:7b \
  --context-length 2048 \
  "Your prompt here"

#Parameter Tuning Guide

#Troubleshooting Common Issues

#Out of Memory Errors

• Use smaller models (7B instead of 13B)
• Reduce context length
• Close other memory-intensive applications
• Set OLLAMA_MAX_MEMORY environment variable

#Slow Response Times

• Check GPU utilization
• Use quantized models (Q4, Q5)
• Keep models loaded with --keep-alive
• Reduce num_predict parameter

#Connection Issues

• Verify Ollama service is running
• Check port 11434 is accessible
• Restart Ollama service
• Check firewall settings