Haozhe Xie

@zhblue 其实是按需下载的，不过是从数据库同步。 不过改成文件系统的确会更好。

So, what's the problem you are facing now? Please provide more details.

@AlphaPav Sorry for the late reply. I don't have time to check this issue these days. But I'm sure that there is nothing wrong with the released dataset. 231792 /...

No. I think the difference may be caused by different compression ratios. You can also generate the ShapeNet dataset from [PCN](https://drive.google.com/drive/folders/1P_W1tz5Q4ZLapUifuOE4rFAZp6L1XTJz) with [this script](https://github.com/hzxie/GRNet/blob/master/utils/lmdb_serializer.py).

@SarahChane98 I cannot report the exact numbers of epochs for each round. For each round, I train several times until the loss does not decrease. Try to fine-tune the network...

@paulwong16 If the reported results are correct, one possible reason why the pretrained model performs worse in the validation and training set is that we choose the best model for...

@paulwong16 Because the pretrained model is best for fitting distribution of the testing set. Instead, the distribution of the training and validation set may be different from the testing set.

@paulwong16 Yes, choosing models from the testing set is not a good option. For the Completion 3D benchmark, the best model is chosen from the validation set. (Because we don't...

@wangyida You can use the Gridding Loss here: https://github.com/hzxie/GRNet/blob/335259235804fa30b0e89e0e257d886687bfb6f3/core/train.py#L113 when fine-tuning the network.

@Lillian9707 In my experiments, the loss will continue to decrease after 20 epochs. Moreover, you need to fine-tune the network with the Gridding Loss.