Unverified50% confidenceFactExact time
Tiny ImageNet is a benchmark dataset derived from Stanford University's CS231N course; it samples 200 classes from ImageNet, compresses resolution to 64×64, and retains 500 training images, 50 validation images, and 50 test images per class.
1
Sources
50%
Confidence
Long-term
Relevance
9/1/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedFiftyOne, developed by Voxel51, supports visualization of multimodal data including images, video, and 3D point clouds, and integrates with training frameworks such as PyTorch and TensorFlow.67% similarUnverified1280×720分辨率在模型训练数据集中出现频率极高，因为大量网页截图和UI录屏素材天然就是这个尺寸63% similarUnverifiedImageMagick 是一套开源图像处理工具集，支持超过 200 种图像格式的读写和转换62% similarUnverified基于扩散模型和生成对抗网络的图像处理技术通常需要海量真实用户图像作为训练数据61% similarUnverifiedImagenette 是 fast.ai 从 ImageNet 中精选的 10 类子集，约 13,000 张图片61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/837577API
curl https://kongchang.com/api/v1/knowledge/claims/837577MCP
get_claim(id=837577)