Fine

Prepare instruction-tuning datasets for LLM fine-tuning. Formats conversations, adds system prompts, balances categories, filters quality. Exports ShareGPT, Alpaca, ChatML, OpenAI formats.. Input: Raw conversations / Q&A pairs. Output: Training dataset (JSONL, Parquet). Python 3.11+. 94% reliable. Use programmatically via API or CLI.

Install via Dejavu

1 Install the Dejavu MCP package:

pip install dejavu-mcp

2 Subscribe at keepingtrack.biz/skills-landing ($6.67/month)

3 Install this skill:

dejavu skill install fine-tuning-data-prep

4 Your AI agent can now use it:

dejavu skill execute fine-tuning-data-prep --input '...'

Pricing

$6.67/month subscription includes 1,000 free credits. This skill uses credits per execution. Subscribe now →

Get Dejavu — $6.67/month