dfaker / embedding-to-png-script Goto Github PK

The text is particularly poorly placed and a bit too big.
It usually shows up on top of the place where faces go
The gray color fade gives it further poor recognition of what is going on under it all.

Can at least the gray be toned down, or simply removed someway, maybe an option for it.
Looking at hundreds of TI embeds that all have a gray pattern over them kind of defeats the point of this.

when I use it for an embedding I made with SD 2.0 (768-v-ema-ckpt) the image gets distorted because it defaults to 512x512. would be cool if you added 768x768 or an option to choose the same resolution as set with the main web ui width & height sliders

Use 中文字 as token shows □□□ in result image

The script usually works great, but on this particular TI embed is throwing:

<tempfile._TemporaryFileWrapper object at 0x00000202F5064070> outputs
100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 18/18 [00:04<00:00, 4.46it/s]
Error completing request
Arguments: ('Style-Glass', '((cartoon)), ((anime)), ((painting)), ((morbid)), ((mutilated)), (((tranny))), (((trans))), johnny, (((trannsexual))), (hermaphrodite), [out of frame], extra fingers, mutated hands, ((poorly drawn hands)), ((poorly drawn face)), (((mutation))), (((deformed))), ((ugly)), blurry, ((bad anatomy)), (((bad proportions))), ((extra limbs)), cloned face, (((disfigured)))', 'None', 'None', 18, 8, False, False, 1, 1, 4, -1.0, -1.0, 0, 0, 0, False, 512, 896, False, 0.7, 2, 'BSRGAN', 30, 896, 512, 3, '', False, <tempfile._TemporaryFileWrapper object at 0x00000202F5064070>, '', 'outputs', 1, 0, 0.25, False, False, False, False, '', 'Illustration', 'svg', True, True, False, 0.5, True, 16, True, 16, False, 1, '', 0, '', True, False, False) {}
Traceback (most recent call last):
File "d:\WORK\conda_envs\automatic\stable-diffusion-webui\modules\call_queue.py", line 45, in f
res = list(func(*args, **kwargs))
File "d:\WORK\conda_envs\automatic\stable-diffusion-webui\modules\call_queue.py", line 28, in f
res = func(*args, **kwargs)
File "D:\WORK\conda_envs\automatic\stable-diffusion-webui\modules\txt2img.py", line 49, in txt2img
processed = modules.scripts.scripts_txt2img.run(p, *args)
File "D:\WORK\conda_envs\automatic\stable-diffusion-webui\modules\scripts.py", line 337, in run
processed = script.run(p, *script_args)
File "D:\WORK\conda_envs\automatic\stable-diffusion-webui\scripts\embedding_to_png.py", line 108, in run
if data.get('step', 0) > 0:
TypeError: '>' not supported between instances of 'NoneType' and 'int'

I am using Auto's web ui: Commit hash: d7aec59c4eb02f723b3d55c6f927a42e97acd679

You can try and reproduce with this file:
https://civitai.com/models/4415/glass-style

Edit: Could be faulty file, but it renders proper so maybe there's something to it.

One thing I noticed is that it seems to embed the number of steps as 1 less than I actually trained it. For example, I trained an embedding to 5000 steps, and the embeddings during training show that. But if I use this script to generate an image embedding from the 5000 step embedding, it only shows 4999 steps on the image. Do you know why that might be?

The original prompt of a shared embedding in png format is too long, I want to change it to something shorter

But it seems I can't just "repack" it by using the png as input for your script

Is there anyway to change the prompt of a png embedding? Change filename is not working