Qwen-Image-Edit-2511-Lightning-4steps_fp8.safetensors

Qwen Image Edit 2511 Lightning 4steps
Qwen Image Edit 2511 Lightning 4steps is a lightweight accelerated variant based on the Qwen Image Edit 2511 basic version, achieving ultra-fast inference with 4-step distilled LoRA while balancing editing quality and interactive experience.
Speed and memory: 4-step sampling, FP8 quantization significantly reduces memory usage.
Model form: distilled LoRA, requires stacking with the Qwen Image Edit 2511 base for use; the quantized version is more suitable for running on mid-to-low-end GPUs.
Capability retention: inherits the semantic editing, text editing, and character consistency of the base, suitable for local modifications, replacements, style transfer, and multi-round iterations.
Model Information
Qwen Image Edit 2511 Lightning 4steps
Qwen Image Edit 2511 Lightning 4steps is a lightweight accelerated variant based on the Qwen Image Edit 2511 basic version, achieving ultra-fast inference with 4-step distilled LoRA while balancing editing quality and interactive experience.
Speed and memory: 4-step sampling, FP8 quantization significantly reduces memory usage.
Model form: distilled LoRA, requires stacking with the Qwen Image Edit 2511 base for use; the quantized version is more suitable for running on mid-to-low-end GPUs.
Capability retention: inherits the semantic editing, text editing, and character consistency of the base, suitable for local modifications, replacements, style transfer, and multi-round iterations.