---
library_name: motius
pipeline_tag: text-to-motion
license: mit
tags:
- human-motion
- two-person-text-to-motion
- intermask
- interhuman
---

# Motius InterMask InterHuman

Inference-only SafeTensors artifact for
[InterMask](https://arxiv.org/abs/2410.10010), adapted from the
[official MIT-licensed repository](https://github.com/gohar-malik/InterMask).
The VQ-VAE and masked Transformer contain no optimizer or scheduler state.

```python
from motius.pipelines.intermask import InterMaskPipeline

pipe = InterMaskPipeline.from_pretrained(
    "ZeyuLing/motius-intermask-interhuman",
    bundle_kwargs={"device": "cuda"},
)
motion = pipe("two people hug", motion_len=120, seed=42)
# (1, 120, 2, 262)
```

See the [full Motius model card](https://github.com/ZeyuLing/Motius/blob/main/docs/model_zoo/intermask.md).