| --- |
| license: apache-2.0 |
| base_model: |
| - Wan-AI/Wan2.2-TI2V-5B |
| pipeline_tag: image-to-video |
| tags: |
| - video-generation |
| - autoregressive-video |
| - video-reasoning |
| - embodied-ai |
| --- |
| |
| # ProAR: Learning Prospective Reasoning with Autoregressive Video Models |
|
|
| ProAR solves reasoning tasks by autoregressively generating a sequence of visual states. It equips autoregressive video models with two complementary components: |
|
|
| - **Outcome Guidance** predicts a future goal frame to guide current-chunk generation. |
| - **Transition Guidance** aligns the current representation with the clean next-chunk representation during training. |
|
|
| ProAR is built upon [Wan2.2-TI2V-5B](https://huggingface.co/Wan-AI/Wan2.2-TI2V-5B). |
|
|
| - **Code:** [github.com/luka-group/ProAR](https://github.com/luka-group/ProAR) |
| - **Evaluation data (for VBVR-10tasks):** [LinghuiShen/ProAR-test](https://huggingface.co/datasets/LinghuiShen/ProAR-test) |
| - **Paper:** [2610.03664](https://arxiv.org/abs/2610.03664) |
| - **Project page:** [luka-group.github.io/ProAR](https://luka-group.github.io/ProAR/) |
|
|
| ## Model Zoo |
|
|
| | Dataset | Model | Checkpoint | Training step | |
| | --- | --- | --- | ---: | |
| | VBVR-10Tasks | AR | [`vbvr-10tasks/ar/model.pt`](./vbvr-10tasks/ar/model.pt) | 10,000 | |
| | VBVR-10Tasks | ProAR | [`vbvr-10tasks/proar/model.pt`](./vbvr-10tasks/proar/model.pt) | 10,000 | |
| | VideoRLVR-3Tasks | AR | [`videorlvr-3tasks/ar/model.pt`](./videorlvr-3tasks/ar/model.pt) | 5,000 | |
| | VideoRLVR-3Tasks | ProAR | [`videorlvr-3tasks/proar/model.pt`](./videorlvr-3tasks/proar/model.pt) | 5,000 | |
| | WorldArena | AR | [`worldarena/ar/model.pt`](./worldarena/ar/model.pt) | 10,000 | |
| | WorldArena | ProAR | [`worldarena/proar/model.pt`](./worldarena/proar/model.pt) | 10,000 | |
|
|
| ## Download |
|
|
| Download all released checkpoints: |
|
|
| ```bash |
| hf download LinghuiShen/ProAR \ |
| --local-dir checkpoints/ProAR |
