亲爱的教授您好：我在下载pccn_bv权重时出现了维度不匹配问题

[wangshuangxing]

数据的坐标，我在您的代码中看到空间坐标是三维的，但是角度分量维度为2，但是下载的权重是3，我该如何操作？

[m-lyon]

This should be fine, the dimensionality of the weights of the hypernetwork do not depend on the dimensionality of the data. Typically, neural network weights only store the input and output channels, and are expanded out to match the shape input shape dimensionality at runtime.

Let me know if you have any issues.