1. Basic Information
Wenxin Yige is an AI-powered art and creative assistance platform launched by Baidu. Leveraging the PaddlePaddle deep learning platform and the Wenxin Big Model, it provides the ability to generate images from text. Users simply enter a Chinese or English description and select a style and frame to generate a high-definition image. The platform is targeted at both casual creators and professional designers, covering a diverse range of subject matter, including illustrations, traditional Chinese posters, realistic fiction, and anime, emphasizing a low-barrier-to-entry creative experience where words become paintings.
2. Product Overview
Wenxin Yige, centered around text-based images, combines style and composition control to help users quickly transform inspiration into creative material. The platform offers templates in multiple styles and commonly used horizontal and vertical ratios, making it suitable for high-frequency scenarios such as social media covers, e-commerce main images, and brand posters. The generation process utilizes a cross-modal model to combine text semantic understanding and image synthesis, supporting continuous experimentation and multiple variations, facilitating the rapid comparison of different solutions. The platform also serves educational and inclusive creative scenarios, providing an intuitive entry point for users lacking a foundation in art, lowering the learning barrier.
3. Core Functions
1. Main functions
- Text to Image
- Simply enter the subject and detailed description to generate an image. It supports Chinese and English prompts and covers common themes such as people, landscapes, products, and decorations.
- Style and composition choices
- It has more than ten built-in styles including traditional Chinese style, oil painting, watercolor, gouache, animation, realism, etc., and provides horizontal, vertical, square and other frame options to facilitate multi-platform adaptation.
- Multiple variations and refinements
- The generated results can continue to generate variants or refine parts to gradually approach the target visual effect.
- Work management and export
- The generated images can be viewed and managed in the personal space, and support high-definition export and re-editing.
- Activities and template resources
- We provide themed activities and style examples from time to time to help creators refer to and reuse common visual schemes.
2. Technical characteristics
- Cross-modal generation
- Based on the Wenxin macro-model and the visual generation sub-model, cross-modal mapping from text to image is completed, improving prompt understanding and detail consistency.
- Paddle frame support
- Relying on PaddlePaddle training and inference optimization strategies, engineering implementation is carried out to ensure the stability and throughput of online services.
- Multilingual and Chinese Enhancement
- It has strong support for fine-grained attributes and style words of Chinese descriptions, and is suitable for precise control in the Chinese creative context.
- Easy-to-use interaction for the general public
- Reduce learning costs through intuitive parameters and style cards, and adapt to desktop and mobile entrances.
- Content and Security Policy
- Built-in basic content review and generation restrictions can intercept and guide non-compliant prompts to ensure compliant use.
4. Pricing and Versions
Wenxin Yige offers free trial quotas and value-added quotas. Newly registered users typically receive a certain amount of electricity for generation. After exceeding this limit, additional quotas can be obtained through top-ups or promotions. Quotas, prices, and benefits may vary across time and region; the actual amount is subject to the platform page. Enterprises or large-scale use cases can integrate solutions related to Baidu Smart Cloud and the Wenxin ecosystem. Capabilities and terms are subject to official specifications.
5. Applicable Scenarios and Target Audience
- New Media and Content Team
- Quickly produce covers and illustrations, adapt to the aspect ratios of multiple platforms, and shorten the image selection and modification cycle.
- E-commerce and brand operations
- Generate theme posters and product scene pictures, and explore styles in batches to unify the visual tone.
- Illustration and visual design
- Conduct style sketches and direction exploration to transform textual ideas into draft visual solutions.
- Education and Training
- Demonstrate the text-to-image process in classrooms and training to lower the threshold for creation for non-art professionals.
- Cultural creativity and event materials
- Combine traditional Chinese style with festival themes to quickly generate event visual elements and improve production efficiency.
6. Frequently Asked Questions
Q: What core capabilities does Wenxin Yige support?
A: It mainly provides the function of generating images from text, with supporting style selection and frame control, and supports variation and refinement of the results to improve controllability.
Q: What is the relationship between Wenxin Yige and Wenxin Da Model?
A: Wenxin Yige realizes text-to-image generation based on the cross-modal capabilities of the Wenxin big model, and completes the engineering deployment of training and inference on the PaddlePaddle framework.
Q: What common styles are available in Wenxin Yige?
A: The platform covers common styles such as traditional Chinese style, oil painting, watercolor, gouache, animation, realism, etc. The specific quantity and name will change with the version update.
Q: How are quotas and fees calculated?
A: The metering is done in units of electricity and other usage units. Registered users can get a certain amount of free quota. After exceeding the quota, they can purchase it on demand or obtain it through activities. The quota and price are subject to the display on the page.
Q: Can the generated work be used commercially?
A: The commercial policies and authorization terms are subject to the latest platform instructions. It is recommended to read and follow the relevant rules before use for brand or commercial activities.