Banner Generation — ZaloAI 2023
Vietnamese marketing copy → production banner · vi2en translation + Qwen LLM keyword expansion + Stable Diffusion generation pipeline
Jan 2023
vi → en
Translation
Stable Diffusion
Generation
Problem
Generate high-quality promotional banner images from raw Vietnamese marketing copy — requiring accurate translation to English, structured keyword extraction for visual cues, and effective prompt construction for diffusion models, all while preserving the original brand tone.
Approach
Engineered an end-to-end text-to-image pipeline for promotional banners — translating Vietnamese marketing copy to English via VinAI vi2en to unlock richer, higher-quality diffusion model prompts.
Amplified prompt quality by using Qwen LLM to extract and expand structured keywords: brand identity, tone, target audience, and visual cues — transforming raw copy into semantically rich generation prompts.
Iterated on Stable Diffusion prompt templates to maximize layout-relevant attributes (style, scene composition, product focus), improving banner visual consistency across diverse marketing briefs.
Results
Delivered a complete pipeline from raw Vietnamese marketing copy to production-ready banner image — no manual prompt engineering required.
Demonstrated that LLM keyword expansion + vi2en translation meaningfully improves Stable Diffusion output quality over direct Vietnamese prompting.
Achieved consistent visual style and layout attributes across diverse marketing briefs through structured prompt templates.
Architecture Overview
Vietnamese copy
VinAI vi2en
Qwen keyword extraction/expansion
Structured prompt
Stable Diffusion
Banner image
How to Run
Install: pip install diffusers transformers torchDownload vi2en model from VinAI HuggingFace hubRun pipeline: python generate_banner.py --input "Vietnamese text here"Output images saved to /outputs/banners/