# GUI-Net-1M

**Repository Path**: hf-datasets/GUI-Net-1M

## Basic Information

- **Project Name**: GUI-Net-1M
- **Description**: Mirror of https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2025-08-11
- **Last Updated**: 2025-08-11

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

---
license: apache-2.0
language:
- en
- zh
---

# Check more details at how to use this dataset at our [**repo**](https://github.com/TongUI-agent/TongUI-agent)
GUI-Net-1M is the dataset we keep running the pipeline introduced from [TongUI paper](https://arxiv.org/abs/2504.12679).
Due to large file size,  we have to split image files into parts. To do the extraction of images, please use the following script:
```bash
#!/bin/bash

# Directory containing the split files
SPLIT_DIR="/mnt/bofeidisk2/tmp/baidu_experience_full/images/split_parts_baidu_experience"
OUTPUT_DIR="merged_files"

# Create output directory if it doesn't exist
mkdir -p "$OUTPUT_DIR"

# Function to merge and unzip files
merge_and_unzip() {
    local base_name=$1
    local output_file="$OUTPUT_DIR/${base_name}.zip"
    
    echo "Processing $base_name..."
    
    # Merge all parts
    cat "$SPLIT_DIR/${base_name}_part"* > "$output_file"
    
    # Unzip the merged file
    echo "Unzipping $output_file..."
    unzip -o "$output_file" -d "$OUTPUT_DIR"
    
    # Remove the zip file after extraction
    rm "$output_file"
}

# Process each main file (0 through 7)
for i in {0..7}; do
    base_name="baidu_experience_full_images_${i}"
    merge_and_unzip "$base_name"
done

echo "All files have been merged and unzipped in the $OUTPUT_DIR directory" 
```
Note: 
  - Change `SPLIT_DIR` into where you download this dataset repo. Change `OUTPUT_DIR` to where you want to un zip those images.
  - For GUI video dataset, change the following lines:
  ```bash
  base_name="baidu_experience_full_images_${i}"
  # change to
  base_name="gui_video_full_images_${i}"
  ```

To run this script, copy paste it into a file `run.sh`, then do 
```bash
bash run.sh
```