stable-diffusion-webui

T-4’s power to do the actual “Canadianizing” of my users’ inputs, and harnessing that power was the easiest part of the project. 2) My project required two art elements — a welcoming natural beaver on the page where the user enters the text, and a robot beaver on a “please wait, the AI is working on it” page. I have never had the slightest artistic ability - I can’t draw a stick. So the ability to use Stable Diffusion on a Google Collab (https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki...), and Bing’s new version of Dall-E (https://www.bing.com/create), to create my art elements was a game-changer. With AI, I was able in a few minutes to create original art that basically satisfied all my requirements for free. 3) Creating the website was a big stretch for me. I’ve never created a web app before, only static websites. Since I’d already started down the path of using the Python version of OpenAi’s API, I decided to build the web app in Python using Flask. At first the process was traditional, googling questions, reading stack overflow, doing a tutorial. GPT-4 was a little helpful here, answering a few questions I had. But as the project progressed it became more complicated. I needed multi-threading so that the website could update while waiting for OpenAI to respond. I needed client side Javascript/AJAX code in the HTML/Jinja2 templates to poll the Python/Flask app and see when it had received the response from OpenAI. Here GPT-4 was a massive accelerator of learning and productivity. I asked it in plain English how to get my code to do this or that and would suggest an approach and some code. I would ask it to explain the code in detail which it did. Having been pointed in a good direction, and understanding what I was doing thanks to GPT-4’s detailed explanations, I was able to get my web app working correctly in just a couple of days. I made many mistakes, and so did GPT-4, but it felt extremely collaborative. It helped me, and I helped it. Every time I suggested it had made a mistake it would try to correct it until together we found a working solution. The end result was an app that’s brought me great pleasure. It’s enabled me to entertain thousands of people, and I’ve gotten tons of positive feedback and thanks. Now that the app is written, it’s really trivial for me to make other versions of it to entertain other communities I’m part of - my local community, audiophiles, headphone fanatics, photographers, etc. The reason it’s so incredibly easy to make other versions is that all I need to do is generate some new art with AI, which takes minutes. And write a new prompt for GPT-4 which also takes minutes. I feel very empowered by these AI tools, and much more confident than before that I can overcome roadblocks with the help of GPT-4.

aybe to the point of feeling like this needs not a mention in the documentation, but it certainly wasn't clear to me. The community around Stable Diffusion is much more inclusive, though. Tools go the extra effort to be easy to use, and documentation for community created models/scripts/tools is so accessible as to be perfectly usable by a non-technical user who is willing to adventure a little bit into the world of hardcore computing by following instructions. Sure, nothing is too polished and you often get the feeling that it's "an ugly thing, but an ugly thing that works", but the point is that it's incredibly accessible. People get to actually use these models to build their stories, fantasy worlds, to work, and things get progressively more impressive as the community builds upon itself (I loved the style of [1] and even effortlessly merged its style with another one in the WebUI, and ControlNet [2] is amazing and gives me ideas for integrating my photography with AI). I think the general interest in creating images is larger than for LLMs with their current limitations (especially in current consumer-available hardware). I do wonder how much this community interest will boost the spaces in the longer run, but right now I can't help but be impressed by the difference in usability and collaborative development between image generative and other types of models. [0] https://github.com/AUTOMATIC1111/stable-diffusion-webui [1] https://civitai.com/models/4998/vivid-watercolors [2] https://github.com/Mikubill/sd-webui-controlnet