metadata
license: mit
dataset_info:
features:
- name: guide_id
dtype: int64
- name: task_title
dtype: string
- name: device_name
dtype: string
- name: difficulty
dtype: string
- name: tools
list: string
- name: time_required_min
dtype: int64
- name: time_required_max
dtype: int64
- name: image_path
dtype: image
- name: text
dtype: string
- name: type
dtype: string
splits:
- name: train
num_bytes: 105039306.48
num_examples: 4495
download_size: 97535095
dataset_size: 105039306.48
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
Electronics Repair Dataset
A comprehensive mini dataset covering smartwatch and other wearable repair and teardown procedures.
Dataset Overview
This dataset contains 4,649 carefully examples
Dataset Structure
Each example contains rich metadata and high-quality repair images:
{
'guide_id': 37170,
'task_title': 'Microsoft Band Wrist Clasp Replacement',
'device_name': 'Microsoft Band',
'difficulty': 'Easy',
'tools': ['T3 Torx Screwdriver'],
'time_required_min': 300,
'time_required_max': 600,
'image_path': <PIL.Image(512, 512)>, # High-quality repair image
'text': 'This is a Microsoft Band Wrist Clasp Replacement guide...',
'type': 'guide_overview' # or 'step_instruction', 'teardown_analysis'
}
Quick Start
from datasets import load_dataset
# Load the complete dataset
dataset = load_dataset("ankithreddy/repairdataset-mini")
# Access examples
example = dataset['train'][0]
image = example['image_path'] # PIL Image object
instruction = example['text'] # Repair instruction text
device = example['device_name'] # Target device
# Filter by category
repair_guides = dataset['train'].filter(lambda x: x['type'] == 'guide_overview')
teardowns = dataset['train'].filter(lambda x: x['type'] == 'teardown_analysis')
steps = dataset['train'].filter(lambda x: x['type'] == 'step_instruction')
This dataset was created using publicly available data from here https://www.ifixit.com/api/2.0/doc for research and educational purposes only. All data belongs to iFixit and respective contributors