#RunPod

Using tori29umai’s LoRA to automatically split facial parts, results from batching 28 images, and a log of running into the limits when attempting finer hair separation

I looked into whether Alibaba Tongyi-MAI's Z-Image can produce stable manga-style character art without breaking the design, using RunPod and ComfyUI.

Configuration for running a Qwen-Image-Layered LoRA that automatically separates facial parts on RunPod. Comparison of RTX 6000 Ada (48GB) and RTX PRO 6000 (96GB).

Hands-on RunPod log for Phr00t AIO Qwen-Image-Edit NSFW v18.1 (28GB). RTX 4090 24GB froze loading and --lowvram didn't help; the FP8 split version needed separate VAE/text encoder. RTX 5090 32GB worked end-to-end. Used it for 3-view reference sheets for a 3D model base mesh.

Setup notes for Qwen-Image-Edit-2511 on RunPod's RTX 4090 ($0.34/hr) using the ComfyUI template. Includes the fal Multiple-Angles LoRA (4 elevations × 8 azimuths × 3 distances) and a per-image cost breakdown that ends up cheaper than buying a 4090.