> For the complete documentation index, see [llms.txt](https://docs.daisy.so/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.daisy.so/core-tools/pods/curating-dataset.md). # Curating dataset **Collect and prepare images that share the same style and aesthetic** Think of your Pod as a unique lens that daisy can switch in and out. The key to successful Pod training is gathering images that all belong to the same visual family - similar style, mood, and artistic approach. *** ## Gathering Your Images ### Where to Find Images
* **Your own library**: Personal photography or artwork * **Stock sources**: Unsplash, Pexels, or other free image sites * **Curated collections**: Any source with consistent style examples ### What to Look For **✅ Do this:** * Choose images with similar artistic styles * Keep consistent color palettes and lighting * Focus on the same overall aesthetic and mood * Stick to similar subjects (all landscapes, all portraits, etc.) **❌ Avoid this:** * Mixing people with scenes or landscapes * Combining different artistic styles (realistic + illustrated) * Using different color themes or lighting conditions * Including images that don't fit the visual pattern *** ## Technical Requirements ### Image Specs * **Format**: JPG or PNG * **Resolution**: Minimum 1000px on one side * **Quality**: High-resolution, not heavily compressed * **Quantity**: 4-15 images (6-8 recommended for best results) ### Preparation Tips * Use image editing tools like Figma to resize and organize * Check resolution before finalizing your set * Export all images in consistent format * Remove any outliers that don't match the style *** **Watch this to learn more:** [Curating Dataset Tutorial](https://youtu.be/HvoYUjgTlxY?si=vdhWpHuDraDXCopf) *** **Ready to Create your Pod?** Once you've prepared your dataset, head to [**Creating a Pod →**](/core-tools/pods/creating-pods.md) to start training.