Anaglyph maker (Stereogram) — v1.0 — v2.0

This ComfyUI workflow turns one 2D image into stereoscopic viewing images, including side-by-side, red-cyan anaglyph, and separate left- and right-eye views.

It uses Qwen-Image and a stereogram LoRA to generate slightly offset left- and right-eye views from the input image.

It stitches the views side by side and merges color channels to make red-cyan anaglyphs, including a black-and-white version.

The workflow accepts one 2D image and saves side-by-side, separate left/right, full-color anaglyph, half-color anaglyph, and black-and-white anaglyph PNGs.

You can choose which formats to save by connecting or disconnecting SaveImage nodes.

The listing recommends about 8–10 GB of VRAM and places model files in ComfyUI’s diffusion_models, loras, text_encoders, and vae folders.

The workflow lists these model files: `Create Stereogram - Qwen Edit 2509_(Make stereogram).safetensors`, `qwen_2.5_vl_7b_fp8_scaled.safetensors`, `qwen_image_edit_2511_bf16.safetensors`, and `qwen_image_vae.safetensors`.

It also lists `Qwen-Image-Edit-2511-Lightning-4steps-V1.0-bf16.safetensors` and `qwen-image-edit-2511-multiple-angles-lora.safetensors`.

Set a fixed SEED for repeatable left/right-view generation, and customize the positive prompt (default: “Make stereogram”) or negative prompt to guide the result.

Adjust Hue, Saturation, and Brightness to tune the anaglyph colors.

The setup text names a 2509 Lightning LoRA, while the workflow lists a 2511 Lightning LoRA; check which filename the workflow expects before setup.

Suggestion · non vérifié

Confirm you have the exact `Create Stereogram - Qwen Edit 2509_(Make stereogram).safetensors` and `qwen-image-edit-2511-multiple-angles-lora.safetensors` files before setup.

Suggestion · non vérifié