Back to datasets
Dataset assetOpen Source CommunityMachine LearningWelding Quality Inspection

TIG_ss304_dataset_full

This dataset is used for welding quality inspection and contains welding images with corresponding labels. Labels are divided into six categories to classify different welding quality issues. The dataset is primarily intended for model training and includes 45,058 training samples.

Source
huggingface
Created
Dec 1, 2024
Updated
Dec 1, 2024
Signals
156 views
Availability
Linked source ready
Overview

Dataset description and usage context

TIG_ss304_dataset_full Dataset Overview

Dataset Information

Features

  • image: Image data
  • label: Label data
    • class_label:
      • names:
        • 0: burn through
        • 1: contamination
        • 2: good weld
        • 3: high travel speed
        • 4: lack of fusion
        • 5: lack of shielding gas

Data Split

  • train:
    • num_bytes: 8,406,037,407.322 bytes
    • num_examples: 45,058 samples

Dataset Size

  • download_size: 9,907,174,988 bytes
  • dataset_size: 8,406,037,407.322 bytes

Configuration

  • config_name: default
    • data_files:
      • split: train
      • path: data/train-*
Need downstream help?

Pair the dataset with AI analysis and content workflows.

Once the source passes your review, move straight into summarization, transformation, report drafting, or presentation generation with the JuheAI toolchain.

Explore AI studio