⚠️ This is a img2img only model — text2img will not work
depth2img preserves the overall form of an image by using auto-generated depth maps.
See screenshots. Often you can get amazing results with very simple prompts.
Tips:
Set Denoising strength = 1 (unless you want to save colors from original image)
512x512 resolution
Best works with volumetric images: 3D renders, photos
Not very good with flat colored 2D art
Description
Original model: https://huggingface.co/stabilityai/stable-diffusion-2-depth
I pruned it to float16 so it doesn't crash a1111 in google colab now.
The yaml is v2-midas-inference: https://raw.githubusercontent.com/Stability-AI/stablediffusion/main/configs/stable-diffusion/v2-midas-inference.yaml
Details
Downloads
662
Platform
CivitAI
Platform Status
Available
Created
1/4/2023
Updated
9/27/2025
Deleted
-
Files
depth2imgPruned_depth2img.ckpt
Mirrors
Huggingface (1 mirrors)
CivitAI (1 mirrors)
depth2imgPruned_depth2img.yaml
Mirrors
CivitAI (1 mirrors)