Training-Free Consistent Text-to-Image Generation
This addresses the challenge for users of text-to-image models who need to portray the same subject consistently across different prompts, offering a training-free solution that is incremental over existing fine-tuning or conditioning methods.
The paper tackles the problem of generating consistent subjects across diverse text prompts in text-to-image models without requiring training or fine-tuning. It introduces ConsiStory, which achieves state-of-the-art performance in subject consistency and text alignment, as demonstrated through comparisons with baselines.
Text-to-image models offer a new level of creative flexibility by allowing users to guide the image generation process through natural language. However, using these models to consistently portray the same subject across diverse prompts remains challenging. Existing approaches fine-tune the model to teach it new words that describe specific user-provided subjects or add image conditioning to the model. These methods require lengthy per-subject optimization or large-scale pre-training. Moreover, they struggle to align generated images with text prompts and face difficulties in portraying multiple subjects. Here, we present ConsiStory, a training-free approach that enables consistent subject generation by sharing the internal activations of the pretrained model. We introduce a subject-driven shared attention block and correspondence-based feature injection to promote subject consistency between images. Additionally, we develop strategies to encourage layout diversity while maintaining subject consistency. We compare ConsiStory to a range of baselines, and demonstrate state-of-the-art performance on subject consistency and text alignment, without requiring a single optimization step. Finally, ConsiStory can naturally extend to multi-subject scenarios, and even enable training-free personalization for common objects.