| license: other | |
| pretty_name: GlobalDiff Inference Data | |
| language: | |
| - en | |
| tags: | |
| - co-speech-gesture-generation | |
| - motion-generation | |
| - diffusion-models | |
| - beat2 | |
| - smpl-x | |
| size_categories: | |
| - n<1K | |
| # GlobalDiff Inference Data | |
| Generated SMPL-X predictions from the official GlobalDiff Speaker 2 release for comparison and visualization without rerunning the full generation pipeline. | |
| - [Paper](https://arxiv.org/abs/2511.10076) · [Hugging Face Paper page](https://huggingface.co/papers/2511.10076) | |
| - [Project page](https://xiangyuezhang.com/GlobalDiff/) · [Code](https://github.com/Xiangyue-Zhang/GlobalDiff) | |
| - [Published version](https://doi.org/10.1609/aaai.v40i15.38281) | |
| - [Model checkpoints](https://huggingface.co/X-Zhang/GlobalDiff) | |
| ## File | |
| `best_pid_2.zip` contains the released generated test results for BEAT2 person 2. | |
| Verify the download with `sha256sum -c SHA256SUMS`. | |
| ## Provenance and terms | |
| The archive contains generated predictions, not BEAT2 ground-truth motion or audio. Users remain responsible for complying with the licenses and terms of [BEAT2](https://huggingface.co/datasets/H-Liu1997/BEAT2), [SMPL-X](https://smpl-x.is.tue.mpg.de/), and the GlobalDiff repository. This card grants no additional rights. | |