Zen Identity Dataset

This dataset contains identity training data for the Zen family of AI models.

Models Covered

• Zen Nano (0.6B): Ultra-efficient edge computing model
• Zen Eco (3B): Balanced performance and efficiency
• Zen Coder (7B): Specialized for code generation
• Zen Omni (14B): Versatile multi-domain model

Dataset Structure

Each example contains:

• instruction: The user's question
• output: The model's response
• model: Which Zen model this example is for
• text: Full conversation format (Human/Assistant)

Usage

from datasets import load_dataset

# Load full dataset
dataset = load_dataset("zenlm/zen-identity")

# Load model-specific data
zen_nano_data = dataset.filter(lambda x: x["model"] == "zen-nano")

Files

• train.jsonl: Complete dataset for all models
• zen-nano_train.jsonl: Zen Nano specific examples
• zen-eco_train.jsonl: Zen Eco specific examples
• zen-coder_train.jsonl: Zen Coder specific examples
• zen-omni_train.jsonl: Zen Omni specific examples

License

Apache 2.0