TimGor1997

> 验证集上： > AP@[IOU0.5|area=all|maxDets=100]=0.482，感觉很不正常。训练时输出的最后几行： > Step: 106999. Epoch: 49/50. Iteration: 1013/2163. Cls loss: 0.01430. Reg loss: 0.2227 > Step: 107499. Epoch: 49/50. Iteration: 1513/2163. Cls loss: 0.00828. Reg loss: 0.1247...

> > Thanks for the quick response! For the Open3D approach do you have code published for this project? > > I just have a simple one for viz. you...

> Thanks @OuyangJunyuan, I would like to have the code for your simple viz. You can send it via mail [[email protected]](mailto:[email protected]). Thank you very much :) hi, could you please...

Dear Junyuan, Glad to hear from you! Thanks for your kindly reply! Your help is greatly appreciated! Best regards, Tim ------------------ 原始邮件 ------------------ 发件人: "xinge008/Cylinder3D" ***@***.***>; 发送时间: 2021年11月6日(星期六) 中午12:54 ***@***.***>; ***@***.******@***.***>; 主题: Re: [xinge008/Cylinder3D]...

Thanks for your reply. I have try to submit the gt.bin, it can get all 1 result, my question is that I have used OpenPCDet to train a model, it...

Thank you for your reply~ But there is a question, how to get the PointPillars and PointRCNN only with 500/125 frames without the remaining 3712-500/125 samples? 

> Those rows are the results for PointPillars/PointRCNN trained with 500/125 human annotations only, no pseudo labels. We use the "500f"/"125f" to denote how many human annotation is required for...

> Yes, that is correct. Thanks. Thank you for the assistance you provided me, answering many of my questions. Looking forward to seeing your new research work! All the Best!

I'm terribly sorry, but I have another question to ask you Why is the number of pseudo_labels out there only 3387 instead of 3769? When you use OpenPCDet's code to...

> We removed labels with too much truncation at [here](https://github.com/Cliu2/MTrans/blob/52e1b81def3ef19b4dded9d1f0faa5008b33e4ea/datasets/kitti_detection.py#L95). The target objects have way too few LiDAR points to generate a good enough pseudo label, and therefore are omitted,...