The datasets are unavailable

[liang3588]

Can you provide the Manifold40 dataset? The link is unavailable.

[zjhJOJO]

I have met the same issue.

you need to dig really hard to their recent paper and find the balanced modelnet40 (https://cg.cs.tsinghua.edu.cn/dataset/subdivnet/datasets/Manifold40.zip). But! you still need to use the preprocessor in their data folder to process the .obj file into the npy files that can be inputed into their model

[zjhJOJO]

you need to dig really hard to their recent paper and find the balanced modelnet40 (https://cg.cs.tsinghua.edu.cn/dataset/subdivnet/datasets/Manifold40.zip). But! you still need to use the preprocessor in their data folder to process the .obj file into the npy files that can be inputed into their model

Thank your for your reply :D

[zjhJOJO]

you need to dig really hard to their recent paper and find the balanced modelnet40 (https://cg.cs.tsinghua.edu.cn/dataset/subdivnet/datasets/Manifold40.zip). But! you still need to use the preprocessor in their data folder to process the .obj file into the npy files that can be inputed into their model

Thank you for your dataset. I still have a question why each mesh in this version has only 500 faces instead of 1024. I remeber that meshes in the initial Manifold40 dataset (not available now) all have 1024 faces.

[JP4CK3T]

you need to dig really hard to their recent paper and find the balanced modelnet40 (https://cg.cs.tsinghua.edu.cn/dataset/subdivnet/datasets/Manifold40.zip). But! you still need to use the preprocessor in their data folder to process the .obj file into the npy files that can be inputed into their model

Does the preprocess.py code in the data folder work well?

For some reason only the ModelNet40_processed directory is created and there is no data in it.