Could you release your text-to-image training data?

[yanjun711]

Hi, I'm trying to reproduce your results, but encountered some issues. The data processing script is not included in the repository.
Moreover, the text-to-image training data is also not included in the meshxl-ShapeNet dataset.

Would it be possible for you to share the complete data processing pipeline or provide access to the text-image paired training data?
This would be crucial for replicating the published results.

[ch3cook-fdu]

We have uploaded the text-to-mesh generation data to huggingface via [link]. We map the object ids with the amazing object captions curated by PointLLM.

[yanjun711]

Thank you for your open source work.