Implementation:Hiyouga LLaMA Factory Alpaca En Demo Data

Knowledge Sources	Hiyouga_LLaMA_Factory
Domains	NLP, Training_Data
Last Updated	2026-02-06 19:00 GMT

Overview

alpaca_en_demo.json provides 500 English instruction-response pairs in the Alpaca format for demonstrating and testing supervised fine-tuning (SFT) workflows in LLaMA Factory.

Description

The file contains a JSON array of 500 records, each following the standard Alpaca data format with three fields: instruction, input, and output. The instructions span a wide variety of tasks including creative writing, classification, summarization, math, and general knowledge questions. The input field is optionally populated when the instruction requires additional context, and the output field contains the expected model response.

This dataset uses the default Alpaca formatting, meaning no special column mappings are required in the dataset registry. It is registered in dataset_info.json under the key "alpaca_en_demo" with only a file_name property.

Usage

This demo dataset is intended for quick testing and validation of SFT training pipelines. Users can reference it by name (alpaca_en_demo) in their training configuration to verify that the data loading, tokenization, and training loop work correctly before switching to larger production datasets.

Code Reference

Source Location

• Repository: Hiyouga_LLaMA_Factory
• File: data/alpaca_en_demo.json

Data Format

[
  {
    "instruction": "Describe a process of making crepes.",
    "input": "",
    "output": "Making crepes is an easy and delicious process! Here are step-by-step instructions..."
  },
  {
    "instruction": "Given the parameters of a triangle, find out its perimeter.",
    "input": "Side 1 = 4\nSide 2 = 6\nSide 3 = 8",
    "output": "The perimeter of a triangle is the sum of the lengths of its sides..."
  }
]

I/O Contract

Schema

Field	Type	Required	Description
instruction	string	Yes	The task or question for the model to respond to
input	string	No	Additional context or input data for the instruction (empty string if unused)
output	string	Yes	The expected model response to the instruction

Dataset Registry Entry

Property	Value
Key	alpaca_en_demo
file_name	alpaca_en_demo.json
formatting	alpaca (default)
Lines	4997
Records	~500

Usage Examples

# Reference the dataset in a LLaMA Factory YAML training config
# examples/train_lora/llama3_lora_sft.yaml
#   dataset: alpaca_en_demo

# Or use it directly via CLI
# llamafactory-cli train \
#     --dataset alpaca_en_demo \
#     --stage sft \
#     --model_name_or_path meta-llama/Llama-2-7b-hf \
#     --output_dir output/sft_demo

# Loading the data manually for inspection
import json

with open("data/alpaca_en_demo.json", "r", encoding="utf-8") as f:
    data = json.load(f)

print(f"Number of samples: {len(data)}")
print(f"First sample instruction: {data[0]['instruction'][:80]}...")
print(f"Has input: {bool(data[0]['input'])}")

Related Pages

• Hiyouga_LLaMA_Factory_Alpaca_Zh_Demo_Data - Chinese version of the Alpaca demo dataset
• Hiyouga_LLaMA_Factory_Dataset_Info_Registry - Central dataset registry that indexes this file
• Hiyouga_LLaMA_Factory_Identity_Data - Identity training data using the same Alpaca format