mlfoundations-cua-dev/easyr1-126k-nores-jedi-fix-synced-ui-vision-manually-labeled-icon-data-from-yt-4MP-aug-jitter

这个数据集是经过增强的图标数据集，基于原始的`easyr1-63k-nores-jedi-fix-synced-ui-vision-manually-labeled-icon-data-from-yt-4MP`数据集，通过对每个样本的坐标进行随机抖动来增加样本的多样性。适用于训练计算机视觉模型，特别是图标识别相关的任务。This dataset is an augmented icon dataset based on the original `easyr1-63k-nores-jedi-fix-synced-ui-vision-manually-labeled-icon-data-from-yt-4MP`, where each samples coordinates are jittered randomly to increase the diversity of the samples. It is suitable for training computer vision models, especially for icon recognition tasks.