vision-utilities
Guide for using vision utilities in speedy_utils, including fast GPU image loading, memory-mapped datasets, and notebook visualization.
npx skills add majiayu000/claude-skill-registry --skill vision-utilities-anhvth-speedy-utils --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
# Vision Utilities Guide This skill provides comprehensive guidance for using the vision utilities in `speedy_utils`. ## When to Use This Skill Use this skill when you need to: - Load images efficiently, leveraging GPU acceleration (NVIDIA DALI) when available. - Create memory-mapped datasets (`ImageMmap`) for extremely fast random access training loops. - Visualize batches of images in Jupyter notebooks with automatic grid layout. - Handle various image formats (numpy, torch, file paths) uniformly. ## Prerequisites - `speedy_utils` installed. - `Pillow` and `numpy` (required). - `matplotlib` (for plotting). - `nvidia-dali-cuda110` or similar (optional, for GPU loading). - `torch` (optional, for tensor support). ## Core Capabilities ### Fast Image Loading (`read_images`) - Tries GPU (DALI) first, falls back to CPU (Pillow). - Supports batch processing and resizing. - Validates images to skip corrupted files. ### Memory-Mapped Datasets (`ImageMmap`, `ImageMmapDynamic`) - **`ImageMmap`**: For fixed-size images. Pre-processes and resizes images once, then stores them in a single binary file for zero-copy access. - **`ImageMmapDynamic`**: For variable-size images. Stores flattened imag
- When to Use This Skill
- Prerequisites
- Core Capabilities
- Fast Image Loading (readimages)
- Memory-Mapped Datasets (ImageMmap, ImageMmapDynamic)
- Notebook Visualization (plotimagesnotebook)
- Usage Examples
- Example 1: Fast Image Loading
- Example 2: Creating a Mmap Dataset
- Example 3: Visualizing Images
- Guidelines
- Limitations
What does the vision-utilities skill do?
Guide for using vision utilities in speedy_utils, including fast GPU image loading, memory-mapped datasets, and notebook visualization.
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill vision-utilities-anhvth-speedy-utils --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.