Sign in to view source links and access this dataset
Description
TIGER-Lab's RationalRewards_DiffusionNFT_TrainData is a reinforcement learning training dataset for text-to-image generation and image editing, associated with the paper 'RationalRewards: Reasoning Rewards Scale Visual Generation Both Training and Test Time'. The dataset was last updated on April 14, 2026. Authors are affiliated with HKUST, University of Waterloo, and Alibaba.
Use Cases
Training RL agents for text-to-image generation based on reasoning rewards.
Fine-tuning diffusion models for image editing based on the described reward mechanism.
Benchmarking RL-based approaches against other visual generation methods.
Studying the impact of reasoning-based rewards on visual generation quality.
Strengths
Dataset is directly associated with a published research paper, providing a clear academic context.
Last updated on 2026-04-14, suggesting recent maintenance.
Limitations
Description metadata is limited; actual data quality requires manual inspection after download.
Column-level documentation is absent; field semantics must be inferred after download.
Row count, file formats, and license are unknown, which may limit suitability assessment.
Provenance
Source
TIGER-Lab, with authors from HKUST, University of Waterloo, and Alibaba.
Collection Method
Created for the 'RationalRewards' research paper, likely involving RL training for visual generation.
Time Range
null
Freshness
Last updated 2026-04-14 06:57:42.
Geography
null
License is unknown; users must verify permissions before use.