AnonymousCervine / depth-image-io-for-SDWebui

An extension to allow managing custom depth inputs to Stable Diffusion depth2img models for the stable-diffusion-webui repo.
72 stars 6 forks source link

Adding tutorial link to use this extension for text artwork generation #6

Open FurkanGozukara opened 1 year ago

FurkanGozukara commented 1 year ago

Hello. First of all thank you so much for this amazing extension.

Just published my tutorial video for how to use this extension to generate epic text

I hope you merge this pull request

Also can you add an option to also export depth mask like in second image?

AnonymousCervine commented 1 year ago

What a lovely video! I'm happy people are using the extension, and coming up with such interesting things to do with it!

I would be happy to host a link to your video here in the documentation—but (many apologies in advance) I'm afraid I'm going to be a bit of a pain in the neck about it (again, sorry!). I hadn't really expected to get submissions to the README here, and I haven't prepared for it. There are legal, technical, and stylistic considerations that I have not worked out in advance.

As a note: If you don't want to deal with any of this, I can just put up a link to your video with my own words and imagery—which would make all of these problems mine and not yours! I would be happy to do so: Everything that follows is predicated on the idea that you might want more say than that in how your work is represented.

Legal:

On the legal side, although it's extremely unlikely to matter, I'd like to get licensing sorted out first—currently the only piece of the repo that has explicit licensing information is the python script itself; that means simple contributions to the readme like this would have ambiguous copyright (which is generally undesirable). Since no such declaration was made prior to you making this pull request, I also would need you to be aware of whatever decision I make (and agree to it, in order to merge the PR—since the agreement normally implicit in making a PR to material with an established license, doesn't apply here).

Now, if I were thinking about this before you made this PR, I'd lean towards just putting the README (and any other documentation) under a CC0 license, but I think that might present potential problems with you putting up an image with what appears(?) to be the likeness of yourself on it (unless you're okay with having that picture of your face in the public domain or equivalent, for anyone to do what they want with). The simplest solution might be instead to use a simpler generated image that demonstrates the technique but does not include a human likeness; one that you're definitely alright with others using freely.

Technical:

At the risk of excessive paranoia, I'd prefer any images in the readme be hosted locally within the repo. This prevents the hypothetical scenario where some malicious contributor changes the content of a linked image unexpectedly, as well as preventing dead image urls (granted that that's all incredibly unlikely in this case).

(This section may be misguided of me, I'm not sure.)

Style:

The proposed presentation is wonderfully flashy, dazzling, and eye-catching, as might be desirable for an advertisement—but the purpose of a readme file is to inform, not to advertise. I would prefer leaning instead towards more context, explanation, and relevant information.

For example:


#### [How To Generate Stunning Epic Text](https://www.youtube.com/watch?v=TBq1bhY8BOc) (youtube)
A video tutorial for using this extension to produce crisp artistically-rendered text. Includes several tips on how to achieve a good-looking result.

<example image goes here>

Drilling down a bit on the details and reasoning:

... Anyway, I hope I haven't upset you with all this nitpicking. I'm keenly aware I'm possibly being too much of a stickler on several fronts here. You have the misfortune of being the first to come knocking on my doorstep with this, when I haven't taken time to mull over which bits actually matter! Many apologies.