Skip to content

Example Modelfiles

The project includes pre-configured Modelfile templates for common use cases. These examples demonstrate best practices and different parameter configurations.

Quick Reference

Example Base Model Temperature Context Use Case
Chatbot llama3.2:1b 0.7 4096 General conversation
Code Assistant llama3.2:1b 0.3 8192 Programming help
Translator llama3.2:1b 0.5 4096 Language translation
Creative Writer llama3.2:1b 1.2 8192 Content creation
Personal Assistant llama3.2:1b 0.6 4096 Task management
TechCorp Support llama3.2:1b 0.3 4096 Customer support

Using Examples

Quick Start

Create a model from any example:

# Interactive selection
make create-model

# Or directly
bash scripts/create-custom-model.sh <model-name> ./models/examples/<example>/Modelfile

# Example:
bash scripts/create-custom-model.sh my-chatbot ./models/examples/chatbot/Modelfile

Test the model: 

make chat
# Select your newly created model

1. Chatbot

Location: models/examples/chatbot/Modelfile

Purpose

General-purpose conversational AI for customer service, Q&A, and casual chat.

Configuration

FROM llama3.2:1b

PARAMETER temperature 0.7
PARAMETER num_ctx 4096
PARAMETER top_p 0.9
PARAMETER repeat_penalty 1.1

SYSTEM """
You are a friendly and helpful AI assistant.
Provide clear, accurate, and concise responses.
Be conversational but professional.
"""

Best For

  • Customer support
  • General Q&A
  • Interactive help systems
  • Educational chatbots

Example Usage

bash scripts/create-custom-model.sh friendly-bot ./models/examples/chatbot/Modelfile
docker compose exec ollama ollama run friendly-bot "How can I track my order?"

Parameter Notes

  • Temperature 0.7: Balanced between consistency and variety
  • Context 4096: Handles medium-length conversations
  • Top P 0.9: Good diversity without being random

2. Code Assistant

Location: models/examples/code-assistant/Modelfile

Purpose

Programming help, code generation, debugging, and technical explanations.

Configuration

FROM llama3.2:1b

PARAMETER temperature 0.3
PARAMETER num_ctx 8192
PARAMETER top_p 0.9
PARAMETER repeat_penalty 1.15

SYSTEM """
You are an expert programming assistant.
- Provide clear, well-commented code
- Explain your reasoning
- Follow best practices and design patterns
- Suggest optimizations when relevant
"""

Best For

  • Code generation
  • Debugging help
  • Code review
  • Technical documentation
  • Algorithm explanations

Example Usage

bash scripts/create-custom-model.sh code-helper ./models/examples/code-assistant/Modelfile
docker compose exec ollama ollama run code-helper "Write a Python function to validate email addresses"

Parameter Notes

  • Temperature 0.3: Deterministic, consistent code generation
  • Context 8192: Can handle large code files
  • Repeat Penalty 1.15: Prevents repetitive code patterns

3. Translator

Location: models/examples/translator/Modelfile

Purpose

Language translation and localization.

Configuration

FROM llama3.2:1b

PARAMETER temperature 0.5
PARAMETER num_ctx 4096
PARAMETER top_p 0.9
PARAMETER repeat_penalty 1.1

SYSTEM """
You are a professional translator.
Translate text accurately while preserving:
- Meaning and intent
- Cultural context
- Tone and formality
- Idiomatic expressions

Provide natural-sounding translations in the target language.
"""

Best For

  • Text translation
  • Localization
  • Language learning
  • Multilingual content creation

Example Usage

bash scripts/create-custom-model.sh translator ./models/examples/translator/Modelfile
docker compose exec ollama ollama run translator "Translate to Spanish: Hello, how are you today?"

Parameter Notes

  • Temperature 0.5: Moderate creativity for natural phrasing
  • Context 4096: Handles paragraphs and documents
  • Top P 0.9: Balanced translation variety

4. Creative Writer

Location: models/examples/creative-writer/Modelfile

Purpose

Creative content generation, storytelling, and ideation.

Configuration

FROM llama3.2:1b

PARAMETER temperature 1.2
PARAMETER num_ctx 8192
PARAMETER top_p 0.95
PARAMETER repeat_penalty 1.2

SYSTEM """
You are a creative writing assistant.
Generate engaging, imaginative, and original content.
Use vivid descriptions, varied vocabulary, and interesting narratives.
Adapt your style to the genre and tone requested.
"""

Best For

  • Story writing
  • Content creation
  • Brainstorming ideas
  • Marketing copy
  • Poetry and creative text

Example Usage

bash scripts/create-custom-model.sh creative ./models/examples/creative-writer/Modelfile
docker compose exec ollama ollama run creative "Write a short story about a robot learning to paint"

Parameter Notes

  • Temperature 1.2: High creativity and variety
  • Context 8192: Long-form content generation
  • Top P 0.95: Wide vocabulary range
  • Repeat Penalty 1.2: Avoids repetitive phrasing

5. Personal Assistant

Location: models/examples/personal-assistant/Modelfile

Purpose

Task management, scheduling, reminders, and personal productivity.

Configuration

FROM llama3.2:1b

PARAMETER temperature 0.6
PARAMETER num_ctx 4096
PARAMETER top_p 0.9
PARAMETER repeat_penalty 1.1

SYSTEM """
You are a personal assistant helping with:
- Task organization and prioritization
- Schedule management
- Reminders and follow-ups
- Information lookup
- Productivity tips

Be efficient, organized, and proactive.
"""

Best For

  • To-do list management
  • Calendar assistance
  • Email drafting
  • Meeting scheduling
  • Personal organization

Example Usage

bash scripts/create-custom-model.sh assistant ./models/examples/personal-assistant/Modelfile
docker compose exec ollama ollama run assistant "Help me plan my day. I have a meeting at 2pm and need to finish a report."

Parameter Notes

  • Temperature 0.6: Consistent but not rigid
  • Context 4096: Tracks tasks and conversations
  • Top P 0.9: Natural conversational flow

6. TechCorp Support

Location: models/examples/techcorp-support/Modelfile

Purpose

Demonstrates few-shot learning with dataset examples for specialized customer support.

Configuration

FROM llama3.2:1b

PARAMETER temperature 0.3
PARAMETER num_ctx 4096
PARAMETER top_p 0.9
PARAMETER repeat_penalty 1.1

SYSTEM """
You are a customer support agent for TechCorp.
Provide accurate, helpful, and professional responses.
"""

# Few-shot examples from dataset
MESSAGE user "How do I reset my password?"
MESSAGE assistant "Click 'Forgot Password' on the login page..."

MESSAGE user "What are your business hours?"
MESSAGE assistant "We're open Monday-Friday, 9am-5pm EST..."

# (additional examples...)

Dataset

The model uses examples from data/training/techcorp-support.jsonl (10 Q&A pairs).

Best For

  • Company-specific support bots
  • FAQ automation
  • Consistent branded responses
  • Knowledge base queries

Example Usage

bash scripts/create-custom-model.sh support-bot ./models/examples/techcorp-support/Modelfile
docker compose exec ollama ollama run support-bot "How do I reset my password?"

Parameter Notes

  • Temperature 0.3: Factual, consistent responses
  • Few-shot Learning: MESSAGE examples train the model
  • Context 4096: Handles support conversations

Learn More

See Dataset Training Example for complete guide on creating custom support bots.

Customizing Examples

All examples can be customized for your needs:

1. Copy the Example

cp -r ./models/examples/chatbot ./models/custom/my-chatbot

2. Edit the Modelfile

nano ./models/custom/my-chatbot/Modelfile

Modify: - Base model: Change FROM line to use different model - Parameters: Adjust temperature, context, etc. - System prompt: Customize behavior and personality - Examples: Add MESSAGE pairs for few-shot learning

3. Create Your Model

bash scripts/create-custom-model.sh my-chatbot ./models/custom/my-chatbot/Modelfile

Parameter Tuning

Quick reference for the values used in the examples above:

Parameter Factual/support Balanced chat Creative
temperature 0.2-0.3 0.6-0.8 1.0-1.2
num_ctx 4096 4096 8192
top_p 0.9 0.9 0.95
repeat_penalty 1.0-1.1 1.1 1.2

For detailed tuning advice, copy-paste presets, and troubleshooting, see the Parameter Guide — the canonical parameter reference.

Testing Your Custom Model

Quick Test

docker compose exec ollama ollama run <model-name> "Test prompt"

Interactive Test

make chat
# Select your model

API Test

curl http://localhost:11434/api/generate -d '{
  "model": "your-model",
  "prompt": "Test prompt",
  "stream": false
}'

Comprehensive Test

Use the quick-test script: 

make quick-test

Next Steps