compare.py

Compare multiple Stable Diffusion models quickly with a contact sheet of multiple images from multiple models with the same prompt

Written and tested on Windows, but should work on Linux with minimal modifications (mostly just the os.system calls)

Requirements

ImageMagik - https://imagemagick.org/script/download.php

• Required for all the post-generation image processing and labeling.

AUTOMATIC1111 stable_diffusion_webui - https://github.com/AUTOMATIC1111/stable-diffusion-webui

• All the image generation uses API calls to AUTOMATIC1111.
• You need to add --api to COMMANDLINE_ARGS in webui-user.bat to enable API access.

Parameters

Run with no arguments or empty arguments, the script will read from 'prompt.txt' and 'prompt_neg.txt' files in the same directory.

Currently supports the following commandline arguments:

options:
  -h, --help            show this help message and exit
  -p PROMPT, --prompt PROMPT
                        Text prompt to run. Leave blank to load ./prompt.txt file
  -n NEGATIVE_PROMPT, --negative-prompt NEGATIVE_PROMPT
                        Negative prompt. Leave blank to load ./prompt_neg.txt file
  -b BATCH_SIZE, --batch-size BATCH_SIZE
                        Number of images to generate per model
  -s STEPS, --steps STEPS
  -d SEED, --seed SEED  Seed to start at, consecutive batch images increment by 1
  -H HIRES, --hires HIRES
                        Enable hi-res fix
  -D DENOISE, --denoise DENOISE
                        De-noising strength
  -c CFG, --cfg CFG     Config Scale

Example

D:\stable-diffusion-webui\outputs>python compare.py -p"female superhero with a cape flying in the sky" -n"ugly, deformed, missing limbs, naked, nude"
[2023-03-11 12:42:44] [1/6] anything-v4.5-pruned.ckpt
[2023-03-11 12:42:44]   🡒 Generating [5] images with [25] steps.
[2023-03-11 12:43:19]   🡒 Saving generated images to disk.
[2023-03-11 12:43:21] [2/6] deliberate_v2.safetensors
[2023-03-11 12:43:28]   🡒 Generating [5] images with [25] steps.
[2023-03-11 12:44:04]   🡒 Saving generated images to disk.
[2023-03-11 12:44:06] [3/6] dreamlike-photoreal-2.0.safetensors
[2023-03-11 12:44:13]   🡒 Generating [5] images with [25] steps.
[2023-03-11 12:44:48]   🡒 Saving generated images to disk.
[2023-03-11 12:44:50] [4/6] DreamShaper_331BakedVae.safetensors
[2023-03-11 12:44:57]   🡒 Generating [5] images with [25] steps.
[2023-03-11 12:45:33]   🡒 Saving generated images to disk.
[2023-03-11 12:45:35] [5/6] NeverendingDreamNED_bakedVae.safetensors
[2023-03-11 12:45:37]   🡒 Generating [5] images with [25] steps.
[2023-03-11 12:46:12]   🡒 Saving generated images to disk.
[2023-03-11 12:46:15] [6/6] moDi-v1-pruned.ckpt
[2023-03-11 12:46:21]   🡒 Generating [5] images with [25] steps.
[2023-03-11 12:46:54]   🡒 Saving generated images to disk.
[2023-03-11 12:46:57]   🡒 Combining all models together and saving generation metadata to text file.
[2023-03-11 12:47:04]   🡒 Completed in 4:20