Getting Started¶
Installation¶
Development Installation¶
Provider Setup¶
Set your API key for your chosen provider:
# OpenAI
export OPENAI_API_KEY="sk-..."
# Anthropic
export ANTHROPIC_API_KEY="sk-ant-..."
# Google Gemini
export GEMINI_API_KEY="..."
Ollama (Local)¶
No API key needed for local models.
Verify Installation¶
Generate Your First Dataset¶
deepfabric generate \
--topic-prompt "Python programming basics" \
--provider openai \
--model gpt-4o \
--num-samples 10 \
--output-save-as dataset.jsonl
This creates a JSONL file with 10 training samples.
Using a Config File¶
For more control, create config.yaml:
topics:
prompt: "Machine learning fundamentals"
mode: tree
depth: 2
degree: 3
generation:
system_prompt: "Generate educational Q&A pairs."
conversation:
type: basic
llm:
provider: openai
model: gpt-4o
output:
system_prompt: "You are a helpful ML tutor."
num_samples: 20
save_as: "ml-dataset.jsonl"
Then run:
Next Steps¶
- Dataset Generation - Types and configuration options
- Tools - Real tool execution for agent datasets
- Training - Using datasets with TRL/Unsloth
- Evaluation - Testing fine-tuned models