Update model card for Tora2: Motion and Appearance Customized Diffusion Transformer for Multi-Entity Video Generation

README.md

@@ -3,6 +3,7 @@ base_model:
 - THUDM/CogVideoX-5b
 language:
 - en
+library_name: diffusers
 license: other
 pipeline_tag: text-to-video
 tags:
@@ -10,22 +11,21 @@ tags:
 - video-generation
 - cogvideox
 - alibaba
-library_name: pytorch
 ---
 
 <div align="center">
 
 <img src="icon.jpg" width="250"/>
 
-<h2><center>[🔥CVPR'25]Tora: Trajectory-oriented Diffusion Transformer for Video Generation</h2>
+<h2><center>[🔥ACM MM'25]Tora2: Motion and Appearance Customized Diffusion Transformer for Multi-Entity Video Generation</h2>
 
 Zhenghao Zhang\*, Junchao Liao\*, Menghao Li, Zuozhuo Dai, Bingxue Qiu, Siyu Zhu, Long Qin, Weizhi Wang
 
 \* equal contribution
 <br>
 
-<a href='https://arxiv.org/abs/2407.21705'><img src='https://img.shields.io/badge/ArXiv-2407.21705-red'></a>
-<a href='https://ali-videoai.github.io/tora_video/'><img src='https://img.shields.io/badge/Project-Page-Blue'></a>
+<a href='https://huggingface.co/papers/2507.05963'><img src='https://img.shields.io/badge/Paper-Tora2-red'></a>
+<a href='https://ali-videoai.github.io/Tora2_page/'><img src='https://img.shields.io/badge/Project-Page-Blue'></a>
 <a href="https://github.com/alibaba/Tora"><img src='https://img.shields.io/badge/Github-Link-orange'></a>
 <a href='https://www.modelscope.cn/studios/xiaoche/Tora'><img src='https://img.shields.io/badge/🤖_ModelScope-ZH_demo-%23654dfc'></a>
 <a href='https://www.modelscope.cn/studios/Alibaba_Research_Intelligence_Computing/Tora_En'><img src='https://img.shields.io/badge/🤖_ModelScope-EN_demo-%23654dfc'></a>
@@ -43,10 +43,12 @@ Zhenghao Zhang\*, Junchao Liao\*, Menghao Li, Zuozhuo Dai, Bingxue Qiu, Siyu Zhu
 
 ## 💡 Abstract
 
-Recent advancements in Diffusion Transformer (DiT) have demonstrated remarkable proficiency in producing high-quality video content. Nonetheless, the potential of transformer-based diffusion models for effectively generating videos with controllable motion remains an area of limited exploration. This paper introduces Tora, the first trajectory-oriented DiT framework that integrates textual, visual, and trajectory conditions concurrently for video generation. Specifically, Tora consists of a Trajectory Extractor (TE), a Spatial-Temporal DiT, and a Motion-guidance Fuser (MGF). The TE encodes arbitrary trajectories into hierarchical spacetime motion patches with a 3D video compression network. The MGF integrates the motion patches into the DiT blocks to generate consistent videos following trajectories. Our design aligns seamlessly with DiT’s scalability, allowing precise control of video content’s dynamics with diverse durations, aspect ratios, and resolutions. Extensive experiments demonstrate Tora’s excellence in achieving high motion fidelity, while also meticulously simulating the movement of physical world.
+Recent advances in diffusion transformer models for motion-guided video generation, such as Tora, have shown significant progress. In this paper, we present Tora2, an enhanced version of Tora, which introduces several design improvements to expand its capabilities in both appearance and motion customization. Specifically, we introduce a decoupled personalization extractor that generates comprehensive personalization embeddings for multiple open-set entities, better preserving fine-grained visual details compared to previous methods. Building on this, we design a gated self-attention mechanism to integrate trajectory, textual description, and visual information for each entity. This innovation significantly reduces misalignment in multimodal conditioning during training. Moreover, we introduce a contrastive loss that jointly optimizes trajectory dynamics and entity consistency through explicit mapping between motion and personalization embeddings. Tora2 is, to our best knowledge, the first method to achieve simultaneous multi-entity customization of appearance and motion for video generation. Experimental results demonstrate that Tora2 achieves competitive performance with state-of-the-art customization methods while providing advanced motion control capabilities, which marks a critical advancement in multi-condition video generation. Project page: this https URL .
 
 ## 📣 Updates
 
+- `2025/07/08` 🔥🔥 Our latest work, [Tora2](https://ali-videoai.github.io/Tora2_page/), has been accepted by ACM MM25. Tora2 builds on Tora with design improvements, enabling enhanced appearance and motion customization for multiple entities.
+- `2025/05/24` We open-sourced a LoRA-finetuned model of [Wan](https://github.com/Wan-Video/Wan2.1). It turns things in the image into fluffy toys. Check this out: https://github.com/alibaba/wan-toy-transform
 - `2025/01/06` 🔥🔥We released Tora Image-to-Video, including inference code and model weights.
 - `2024/12/13` SageAttention2 and model compilation are supported in diffusers version. Tested on the A10, these approaches speed up every inference step by approximately 52%, except for the first step.
 - `2024/12/09` 🔥🔥Diffusers version of Tora and the corresponding model weights are released. Inference VRAM requirements are reduced to around 5 GiB. Please refer to [this](diffusers-version/README.md) for details.
@@ -156,6 +158,91 @@ git clone https://www.modelscope.cn/xiaoche/Tora.git
     - T5: [text_encoder](https://huggingface.co/THUDM/CogVideoX-2b/tree/main/text_encoder), [tokenizer](https://huggingface.co/THUDM/CogVideoX-2b/tree/main/tokenizer)
 - Tora t2v model weights: [Link](https://cloudbook-public-daily.oss-cn-hangzhou.aliyuncs.com/Tora_t2v/mp_rank_00_model_states.pt). Downloading this weight requires following the [CogVideoX License](CogVideoX_LICENSE).
 
+## 🔄 Inference
+
+### Text to Video
+It requires around 30 GiB GPU memory tested on NVIDIA A100.
+
+```bash
+cd sat
+PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True torchrun --standalone --nproc_per_node=$N_GPU sample_video.py --base configs/tora/model/cogvideox_5b_tora.yaml configs/tora/inference_sparse.yaml --load ckpts/tora/t2v --output-dir samples --point_path trajs/coaster.txt --input-file assets/text/t2v/examples.txt
+```
+
+You can change the `--input-file` and `--point_path` to your own prompts and trajectory points files. Please note that the trajectory is drawn on a 256x256 canvas.
+
+Replace `$N_GPU` with the number of GPUs you want to use.
+
+### Image to Video
+
+```bash
+cd sat
+PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True torchrun --standalone --nproc_per_node=$N_GPU sample_video.py --base configs/tora/model/cogvideox_5b_tora_i2v.yaml configs/tora/inference_sparse.yaml --load ckpts/tora/i2v --output-dir samples --point_path trajs/sawtooth.txt --input-file assets/text/i2v/examples.txt --img_dir assets/images --image2video
+```
+
+The first frame images should be placed in the `--img_dir`. The names of these images should be specified in the corresponding text prompt in `--input-file`, seperated by `@@`.
+
+### Recommendations for Text Prompts
+
+For text prompts, we highly recommend using GPT-4 to enhance the details. Simple prompts may negatively impact both visual quality and motion control effectiveness.
+
+You can refer to the following resources for guidance:
+
+- [CogVideoX Documentation](https://github.com/THUDM/CogVideo/blob/main/inference/convert_demo.py)
+- [OpenSora Scripts](https://github.com/hpcaitech/Open-Sora/blob/main/scripts/inference.py)
+
+## 🖥️ Gradio Demo
+
+Usage:
+
+```bash
+cd sat
+python app.py --load ckpts/tora/t2v
+```
+
+## 🧠 Training
+
+### Data Preparation
+
+Following this guide https://github.com/THUDM/CogVideo/blob/main/sat/README.md#preparing-the-dataset, structure the datasets as follows:
+
+```
+.
+├── labels
+│   ├── 1.txt
+│   ├── 2.txt
+│   ├── ...
+└── videos
+    ├── 1.mp4
+    ├── 2.mp4
+    ├── ...
+```
+
+Training data examples are in `sat/training_examples`
+
+### Text to Video
+
+It requires around 60 GiB GPU memory tested on NVIDIA A100.
+
+Replace `$N_GPU` with the number of GPUs you want to use.
+
+- Stage 1
+
+```bash
+PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True torchrun --standalone --nproc_per_node=$N_GPU train_video.py --base configs/tora/model/cogvideox_5b_tora.yaml configs/tora/train_dense.yaml --experiment-name "t2v-stage1"
+```
+
+- Stage 2
+
+```bash
+PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True torchrun --standalone --nproc_per_node=$N_GPU train_video.py --base configs/tora/model/cogvideox_5b_tora.yaml configs/tora/train_sparse.yaml --experiment-name "t2v-stage2"
+```
+
+## 🎯 Troubleshooting
+
+### 1. ValueError: Non-consecutive added token...
+
+Upgrade the transformers package to 4.44.2. See [this](https://github.com/THUDM/CogVideo/issues/213) issue.
+
 ## 🤝 Acknowledgements
 
 We would like to express our gratitude to the following open-source projects that have been instrumental in the development of our project:
@@ -174,13 +261,10 @@ Special thanks to the contributors of these libraries for their hard work and de
 ## 📚 Citation
 
 ```bibtex
-@misc{zhang2024toratrajectoryorienteddiffusiontransformer,
-      title={Tora: Trajectory-oriented Diffusion Transformer for Video Generation},
+@article{zhang2025tora2,
+      title={Tora2: Motion and Appearance Customized Diffusion Transformer for Multi-Entity Video Generation},
       author={Zhenghao Zhang and Junchao Liao and Menghao Li and Zuozhuo Dai and Bingxue Qiu and Siyu Zhu and Long Qin and Weizhi Wang},
-      year={2024},
-      eprint={2407.21705},
-      archivePrefix={arXiv},
-      primaryClass={cs.CV},
-      url={https://arxiv.org/abs/2407.21705},
+      journal={ACM Multimedia (MM)},
+      year={2025}
 }
 ```