clawCa

Results 10 comments of clawCa

> 应该不是环境问题,这可能是因为没有用多卡训练。 跑通了,是s3dis处理后的路径不对,导致没有正确读到数据。关于论文有2个细节不太明白,想请教您 ![image](https://user-images.githubusercontent.com/71144235/178760297-b923fcd6-09b1-4489-9729-9e428a2147a7.png) 这里的( N_(s+1) ,k )是降采样后的 N_(s+1)的点周围k个原始点的index吗? 还是获取为聚合k个点的信息为k维的特征向量?还有就是这里的k和c是一样的值,上下拼接做max pooling吗? ![image](https://user-images.githubusercontent.com/71144235/178763003-d96691e9-b17d-41f7-85b7-8bb8f904dfaa.png) 第二个是关于table t_x t_y t_z 是如何建立的? 是把窗口中的点什么信息进行self-attention获得的table吗?

> 嗯……我没有理解你的问题。 > > 你是指在看 pdf、还是某些论坛网站?具体是在哪个 pdf 或者网站里,chrome 和 edge 的表现是不一样的?你指的是划词还是截图识别? 图片中,一个是句子被换行后翻译不完整,另一个是希望能加个开关翻译后每一句话都另起一行开始,阅读清晰一些 图片上传会失败,图片在下面 ![img](https://user-images.githubusercontent.com/71144235/161012521-5a51b725-8438-4828-85c4-f5c30fa1d94f.jpg)

![img](https://user-images.githubusercontent.com/71144235/161012330-c24ba210-7749-41e7-bca4-b2d6d1ceb619.jpg) 发图失败了,重发图片

> 方便提供给我这个 pdf 文件我试试吗?我需要先在我这里重现你的问题才能解决这个问题 https://arxiv.org/pdf/2203.11139 这个链接,不知道怎么传,就发链接给你吧

这是每class的名称,你按照类别写入这个文件中就好了

This issue is over due to a mismatch between voxel size and point cloud data; New Question Q1: How to implement the learnable position encoding table in the paper? I...

> Hi! May I please ask you to share the versions of the modules that are working for you? I am getting the error especially for hydra. It seems that...

> > Thank you for open sourcing your code! I trained the model on my own dataset. In my segmentation scenario, I have a class representing over 30 potted plants....

> Thank you for your issues,may I know how your dataset is made, and how to use it? You need to modify the pre-processing code provided by the author based...

Encountered the same problem. It seems like there was an error in the dataset preprocessing. Did you use the conversion script from mmdetection3d for the S3DIS dataset? In that script,...