MeshFlow: Processed 3D Mesh Dataset for Equivariant Flow Matching

A processed dataset of 3D meshes for training the MeshFlow model, which uses equivariant flow matching for mesh generation. The dataset is hosted by author qsun2001 and was last updated on June 16, 2026. It accompanies a research paper and includes pretrained model checkpoints.

Use Cases

• Training generative models for 3D mesh synthesis based on the described equivariant flow matching method.
• Benchmarking mesh generation techniques using the provided processed mesh data.
• Fine-tuning pretrained models for specific 3D shape generation tasks based on the available checkpoints.

Strengths

• Includes pretrained model checkpoints, enabling immediate use of the MeshFlow method.
• Associated with a published research paper and a live demo, providing context and validation.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and dataset size are unknown, which may limit suitability assessment.

Provenance

Source

Author qsun2001 via Hugging Face.

Collection Method

Processed mesh data, likely derived from other 3D datasets, for use with the MeshFlow model.

Freshness

Last updated 2026-06-16 07:49:00; freshness should be verified.