The hardware requirements for Qwen-Image are listed below:
- Recommended ConfigurationsThe first run takes about 86% of video memory, the second run takes about 71 seconds, and the first run takes about 86% of video memory, the second run takes about 71 seconds. The first run takes about 86% of video memory and takes 94 seconds; the second run takes about 71 seconds.
- minimum configuration: CPUs can also run it, but it is slower and not suitable for high load tasks.
- Low memory optimization: Supports low video memory inference (4GB VRAM) and FP8 quantization through the ModelScope platform for resource-constrained users.
In addition, when deploying multi-GPU services, the number of GPUs, task queue size, and timeout can be configured via environment variables to support high concurrency requirements.
This answer comes from the articleQwen-Image: an AI tool for generating high-fidelity images with accurate text renderingThe
May not be reproduced without permission:AI productivity tools " What are the hardware requirements for Qwen-Image?
English 
简体中文 
日本語 
Deutsch 
Português do Brasil