learnuser1

等你开源的时候已经有很多复现的版本了吧

> ``` > 您好，我自己用完整的nuScenes数据集训练，显卡是两张3090， BATCH_SIZE_PER_GPU=8，用的是‘cbgs_voxel0075_voxelnext.yaml’配置文件。训练时间稳定的话大概是200个小时，这是正常的吗。 > ``` > 你好，这个自己训练的话效果跟原论文比的话怎么样。然后这个模型跟centerpoint比的话要好很多嘛

> > > ``` > > > 您好，我自己用完整的nuScenes数据集训练，显卡是两张3090， BATCH_SIZE_PER_GPU=8，用的是‘cbgs_voxel0075_voxelnext.yaml’配置文件。训练时间稳定的话大概是200个小时，这是正常的吗。 > > > ``` > > > > > > > > > > > > > 你好，这个自己训练的话效果跟原论文比的话怎么样。然后这个模型跟centerpoint比的话要好很多嘛 > > 我的机器跑的很慢，20个epoch还没跑完。但是我测试了第10个epoch，mAP已经到了59.32了。CenterPoint我没跑过，但是论文里面写的是CenterPoint的mAP为...

> voxelnext_nuscenes_kernel1.pth https://pan.baidu.com/s/1SSvIIQpafZmhFXoEI2AjYQ 提取码: x4d9 > > VoxelNeXt_Argo2.pth https://pan.baidu.com/s/1Fw-uWVl2dCsx_HvCr-NmmQ 提取码: kt9p 非常感谢，好像我用预训练模型在test.py测试的话性能跟论文结果差挺多的，不知道什么原因

> 请问你测试用的脚本，和结果是多少？ 我用的python test.py --cfg_file ${cbgs_voxel0075_voxelnext_doubleflip.yaml} --batch_size ${BATCH_SIZE} --ckpt ${CKPT}，是直接在OpenPCDet上的跑的，准备再用自己的数据集来跑一下推理，不过不知道怎么处理数据了。。。

> Hello author, how do you visualize the results? Can you provide the corresponding source code? Hello, can you provide visual code? Thank you very much

I tested it on my own computer and it didn't work out too well.

> When I was following your guidance to use dwpose to extract the pose video, I find the results are jittering. However, it seems your dwpose video are stable, what...

> We tried rescaling through the external rectangle of the kps common to ref and motion, and tested in **Moore-AnimateAnyone**, but the effect was not significantly improved. I tried scaling...

> @TZYSJTU I have used simple average filtering, and it was little bit simple and effective for me > > @ailingzengzzz Thanks for great work and help! Does simple average...