Calandra Dataset

该数据集名为Calandra，包含了用于预测抓握结果的样本，其中非重叠的样本被划分为训练集和测试集。该数据集还包括来自单个触觉传感器的触觉感知数据，用于评估对比学习方法的性能。在规模上，训练集包含6,000个样本，测试集包含3,000个样本。该数据集的任务是抓握结果预测。

The dataset named Calandra encompasses samples for grasp outcome prediction, with non-overlapping samples partitioned into training and test sets. It additionally incorporates tactile perception data collected from a single tactile sensor, which is utilized to assess the performance of contrastive learning approaches. Regarding its scale, the training set consists of 6,000 samples, and the test set contains 3,000 samples. The task of this dataset is grasp outcome prediction.

搜集汇总

背景与挑战

背景概述

Calandra数据集专注于抓握结果预测任务，包含来自单个触觉传感器的感知数据，用于评估对比学习方法。数据集规模为6,000个训练样本和3,000个测试样本，样本非重叠划分，适用于机器学习模型训练与性能测试。