This is a text-to-image application based on ERNIE-Image.

After the user inputs a prompt, the workflow first creates an empty latent based on the set aspect ratio, combines positive and negative prompts into the sampler for generation, and finally outputs the image via VAE decoding. Unlike standard text-to-image, this pipeline incorporates an entire Prompt Enhancement structure. It automatically expands a user's brief description into a richer, more detailed visual description before sending it to the main model for generation. Therefore, it is ideal for users who do not know how to write long prompts but still want to directly generate high-quality images.

This workflow is suitable for:

Poster generation
Cover image generation
Illustration generation
Product display image generation
Structured infographic, fashion, and stylized single-image generation

Recommended usage:

Directly input a brief scene description
Enable prompt enhancement to let the system automatically add details
If you already have a complete prompt, you can disable enhancement and generate directly using the original prompt
Best suited for single-image generation; not suitable for complex narrative prompts

Recommended prompt example:

A young woman standing on a neon-lit night street, cinematic lighting, high-end fashion photography style, exquisite visuals, rich details

🎁Claim your RH Coins first, then try out the workflow! Go to the top-right avatar → Invitation Code → Enter [rh-v1111] to instantly get 1,000 RH Coins. Log in daily to claim another 100 coins~ For more ComfyUI workflows, tutorials, and gameplay, follow our official WeChat account (AIKSK), and stay tuned for simultaneous updates on Douyin / Bilibili / Xiaohongshu / YouTube (AI-KSK) ✨

Ernie generates images