qwen_image_play

Create still images using Huggingface diffusers library and latest Qwen Image model

"A couple strolling in a dimly lit dark night, digital painting, artstation, concept art, Sergey Kolesov , Breathtaking, 8k resolution, extremely detailed, beautiful, establishing shot, artistic, hyperrealistic, octane render, cinematic lighting, dramatic lighting, masterpiece, light brazen, extremely detailed and beautiful face"

Hardware Configuration tested

• Intel/AMD system with at least 24 cores
• 256 GB DDR5
• NVIDIA 5090 w/ 32GB VRAM
• 80GB free storage space for models

Prerequisites

System setup with:

• kernel drivers for NVIDIA GPU
• docker with NVIDIA-Runtime support
• 'make' installed with your package manager (e.g. apt install make)

Steps for setup and execution

1. Build docker container

$ make image

2. Run sample to generate three different images from prompts under scripts/prompts.txt

$ make run

On NVIDIA 5090 card, it takes about 1 minute to generate each image. They will be located as 0.png, 1.png, and 2.png in the 'output_folder' folder

References

• https://prompthero.com
• https://arxiv.org/abs/2508.02324
• https://huggingface.co/Qwen/Qwen-Image
• https://huggingface.co/DFloat11/Qwen-Image-DF11