spaicom-lab/latents-cifar10

该数据集是针对cifar10数据集的timm模型预计算嵌入，包含了多个模型的训练和测试数据集。每个模型对应一个数据集配置，数据以Parquet格式存储，并可通过load_dataset函数进行加载。数据集包含字段有example_id、label、model_name和embedding，其中embedding的大小随模型不同而变化。

This dataset contains precomputed embeddings for timm models on the cifar10 dataset, including training and testing datasets for multiple models. Each model corresponds to a dataset configuration, with data stored in Parquet format and loadable via the load_dataset function. The dataset includes fields such as example_id, label, model_name, and embedding, where the size of embedding varies with different models.