Hey all, I think this is going to be a bit lengthy post but before that, me hopes you enjoys this set. More of a digital painting style. Theme : MadMax Pawamo and Lysandre (and some others) were asking about the creation process, so I'm just going to put some ideas down. Since I have also just begun this exploration I'm not even going to call this a tutorial ok hehe...just sharing whatever little experience I've had; what worked and some simple tips etc... Plus as some of you may know I don't have lofty goals in creating art - pretty girls in pretty outfits would roughly sum up my contribution to the galleries :)), so I can't say for sure these ideas will work on other genres/subject matters. I'll try to address in point form to stay on track because I do have a tendency to babble :p In the end I might or might not have answered you comment. If not, let me know and I'll try me best to give a better answer. 1. Pamawo : I'm using text to image generator, so there is no image to start from (except the one in my head :D) There are also image to image art generators but I haven't explored them yet so I can't give any pointers or tips for now. Roughly the process for text to image is you start with some basic prompts, then based on the image generated, you add more prompts and re-generate, and so on until you're happy with the result. Yes in the end your moulinette might look like some recipe for world domination hehe but what's important is the resulting image :) For those just starting out, try to think of generalities, instead of trying to produce something strictly like you imagined it (something you would try to do in poser or daz), because AI art is something else. Since we're just starting, such a goal might not be so easily achieved. However, as you get better you will learn to give better prompts and directions. 2. Lysandre : I'm using the paid version of Mage.Space which costs $15 a month for unlimited images (and NSFW :-)) For those who are more advanced, you might want to check out point no 4 for some ideas of what you can't do with that platform that you are used to doing elsewhere, to help you decide. The good thing about Mage.Space (I'm not affiliated to it in anyway! XD), is that for beginners like me, it has lots of examples created by the model/style makers themselves, so they are not 'whitewashed' prompts so to speak. I can't vouch for other platforms of course, but if you've been in the situation where you like an AI art and tried to generate based on the prompts they supplied and get p**p, you know what I mean. Granted, no guarantees that a correct/genuine prompt will give the same result, but sometimes you know because the results are, well...too p**py hehe... For beginners : Models/styles are basically a certain way that the AI has been trained so that their results look better/consistent (That's just my own simple definition) However, you will still more often than not get 6-7 fingers on each hand :p Paid subscribers get around 40 finetuned models/styles that include realistic, anime, synthwave punk, portraits, cute, disney/pixar etc. If you don't want NSFW you can turn it off in the settings (why??? lol j/k) I've tried a good number of these models and they are really excellent. Free version has only 2 models, but you can still generate unlimited images. 3. As for prompts, what I did was look at the examples provided, find one that I like and basically hit 'Rerun' and it will generate a new art based on those prompts, which look great and yes it's considered your art (yay) Congrats! Then what I did was add a few prompts and rerun, and kept repeating the process until I got something I like, and also understood how I got there. The advantages of having these examples are that they actually work, so it kinda keeps you going, rather than striking out purely on your own and getting discouraged by the results and saying AI is not for you. And secondly, these examples include prompts that you probably won't have thought of using, or even know about. So they are really eye openers and can expand your choice of prompts (and by extension help you create your own advanced prompts). Also, I would suggest you do some 'reverse engineering' by deleting some prompts and seeing how it affects the results. 4. Some shortcomings : For advanced users, please take note : a. You can't generate multiple images at one go (yeah, bummer lol) b. You can't add more models, e.g those you find on civitai etc c. I haven't seen any LoRA models yet. d. You can't choose the sampler and I'm not sure what it's defaulted at. e. You can set aspect ratio, steps, guidance scale, seed, and negative prompts That's about it for now. Hope it helps somewhat. If I was unclear let me know and I'll try to do better at explaining it. Thanks all, and as always me hopes you enjoys :D