--- dataset_info: - config_name: train features: - name: video_path dtype: string - name: internal_id dtype: string - name: prompt dtype: string - name: url dtype: string - name: annotation struct: - name: alignment dtype: int64 range: [1,5] - name: composition dtype: int64 range: [1,3] - name: focus dtype: int64 range: [1,3] - name: camera movement dtype: int64 range: [1,3] - name: color dtype: int64 range: [1,5] - name: lighting accurate dtype: int64 range: [1,4] - name: lighting aes dtype: int64 range: [1,5] - name: shape at beginning dtype: int64 range: [0,3] - name: shape throughout dtype: int64 range: [0,4] - name: object motion dynamic dtype: int64 range: [1,5] - name: camera motion dynamic dtype: int64 range: [1,5] - name: movement smoothness dtype: int64 range: [0,4] - name: movement reality dtype: int64 range: [0,4] - name: clear dtype: int64 range: [1,5] - name: image quality stability dtype: int64 range: [1,5] - name: camera stability dtype: int64 range: [1,3] - name: detail refinement dtype: int64 range: [1,5] - name: letters dtype: int64 range: [1,4] - name: physics law dtype: int64 range: [1,5] - name: unsafe type dtype: int64 range: [1,5] - name: safety dtype: int64 range: [1,5] - name: meta_result sequence: dtype: int64 - name: meta_mask sequence: dtype: int64 splits: - name: train num_examples: 40743 - config_name: regression features: - name: internal_id dtype: string - name: prompt dtype: string - name: standard_answer dtype: string - name: video1_path dtype: string - name: video2_path dtype: string splits: - name: regression num_examples: 1795 - config_name: test features: - name: internal_id dtype: string - name: prompt dtype: string - name: standard_answer dtype: string - name: video1_path dtype: string - name: video2_path dtype: string splits: - name: test num_examples: 1000 configs: - config_name: train data_files: - split: train path: train/*.parquet - config_name: regression data_files: - split: regression path: regression/*.parquet - config_name: test data_files: - split: test path: test/*.parquet license: apache-2.0 --- # VisionRewardDB-Video This dataset is a comprehensive collection of video evaluation data designed for multi-dimensional quality assessment of AI-generated videos. It encompasses annotations across 21 diverse aspects, including text-to-video consistency, aesthetic quality, motion dynamics, physical realism, and technical specifications. 🌟✨ [**Github Repository**](https://github.com/THUDM/VisionReward) 🔗 The dataset is structured to facilitate both model training and standardized evaluation: - `Train`: A primary training set with detailed multi-dimensional annotations - `Regression`: A regression set with paired preference data - `Test`: A video preference test set for standardized performance evaluation This holistic approach enables the development and validation of sophisticated video quality assessment models that can evaluate AI-generated videos across multiple critical dimensions, moving beyond simple aesthetic judgments to encompass technical accuracy, semantic consistency, and dynamic performance. ## Annotation Detail Each video in the dataset is annotated with the following attributes:
Dimension Attributes
Alignment Alignment
Composition Composition
Quality Color; Lighting Accurate; Lighting Aes; Clear
Fidelity Detail Refinement; Movement Reality; Letters
Safety Safety
Stability Movement Smoothness; Image Quality Stability; Focus; Camera Movement; Camera Stability
Preservation Shape at Beginning; Shape throughout
Dynamic Object Motion dynamic; Camera Motion dynamic
Physics Physics Law
### Example: Camera Stability - **3:** Very stable - **2:** Slight shake - **1:** Heavy shake - Note: When annotations are missing, the corresponding value will be set to **-1**. For more detailed annotation guidelines(such as the meanings of different scores and annotation rules), please refer to: - [annotation_detail](https://flame-spaghetti-eb9.notion.site/VisioinReward-Video-Annotation-Detail-196a0162280e8077b1acef109b3810ff?pvs=4) - [annotation_detail_zh](https://flame-spaghetti-eb9.notion.site/VisionReward-Video-196a0162280e80e7806af42fc5808c99?pvs=4) ## Additional Feature Detail The dataset includes two special features: `annotation` and `meta_result`. ### Annotation The `annotation` feature contains scores across 21 different dimensions of video assessment, with each dimension having its own scoring criteria as detailed above. ### Meta Result The `meta_result` feature transforms multi-choice questions into a series of binary judgments. For example, for the `Camera Stability` dimension: | Score | Is the camera very stable? | Is the camera not unstable? | |-------|--------------------------|---------------------------| | 3 | 1 | 1 | | 2 | 0 | 1 | | 1 | 0 | 0 | - note: When the corresponding meta_result is -1 (It means missing annotation), the binary judgment should be excluded from consideration Each element in the binary array represents a yes/no answer to a specific aspect of the assessment. For detailed questions corresponding to these binary judgments, please refer to the meta_qa_en.txt file. ### Meta Mask The `meta_mask` feature is used for balanced sampling during model training: - Elements with value 1 indicate that the corresponding binary judgment was used in training - Elements with value 0 indicate that the corresponding binary judgment was ignored during training ## Data Processing ```bash cd videos tar -xvzf train.tar.gz tar -xvzf regression.tar.gz tar -xvzf test.tar.gz ``` We provide `extract.py` for processing the `train` dataset into JSONL format. The script can optionally extract the balanced positive/negative QA pairs used in VisionReward training by processing `meta_result` and `meta_mask` fields. ```bash python extract.py ``` ## Citation Information ``` @misc{xu2024visionrewardfinegrainedmultidimensionalhuman, title={VisionReward: Fine-Grained Multi-Dimensional Human Preference Learning for Image and Video Generation}, author={Jiazheng Xu and Yu Huang and Jiale Cheng and Yuanming Yang and Jiajun Xu and Yuan Wang and Wenbo Duan and Shen Yang and Qunlin Jin and Shurun Li and Jiayan Teng and Zhuoyi Yang and Wendi Zheng and Xiao Liu and Ming Ding and Xiaohan Zhang and Xiaotao Gu and Shiyu Huang and Minlie Huang and Jie Tang and Yuxiao Dong}, year={2024}, eprint={2412.21059}, archivePrefix={arXiv}, primaryClass={cs.CV}, url={https://arxiv.org/abs/2412.21059}, } ```