Stable Diffusion v2.0 and Prompts

Stable Diffusion v2.0 and Prompts

Stable Diffusion v2.0 and prompts

The update brings improvements to SD’s text-to-image diffusion models, includes a powerful image upscaler, updates its Inpainting model, and more.

Also it appears to have made massive updates to the prompting system comparing Stable diffusion v1.4 to v2 gives pretty vast differences. They produced a prompts guide which was online but seems to have disappeared?

But irritatingly it can’t be downloaded as a .pdf, me being a massive nerd said f**k that and used python to pull the svgs and created a pdf.

Here are a few examples of how different things can get

So here are a few examples with the relevant seeds and prompts to show you how dramatically things have changed.   Also, it should be mentioned a lot of celebrities have been removed from the model, NSFW images and famous artists like Rutkowski have been purged due to complaints of their style being essentially stolen by the machine.

What has been removed from Stable Diffusion’s training data, though, is nude and pornographic images. AI image generators are already being used to generate NSFW output, including both photorealistic and anime-style pictures. However, these models can also be used to generate NSFW imagery resembling specific individuals (known as non-consensual pornography) and images of child abuse.

Of course, there are a lot of angry Incels now wailing about censorship. This is bullshit, this stuff is open-source. No doubt a horde of horny teenagers are using PornHub and other sites to train NSFW models. 

Prompt: Gandalf, d & d, fantasy, intricate, elegant, highly detailed, digital painting, artstation, concept art, matte, sharp focus, illustration, hearthstone, art by artgerm and greg rutkowski and alphonse mucha
Negative prompt: cartoon, 3d, ugly face, (disfigured), (bad art), (deformed), (poorly drawn), (extra limbs), strange colours, blurry, boring, sketch, lacklustre, repetitive, cropped, hands

Steps: 60, Sampler: DDIM, CFG scale: 12, Seed: 1940895508, Face restoration: GFPGAN, Size: 512×512, Model hash: 7460a6fa, Batch size: 6, Batch pos: 0

gandalf, d & d, fantasy, intricate, elegant, highly detailed, digital painting, artstation, concept art, matte, sharp focus, illustration, hearthstone, art by artgerm and greg rutkowski and alphonse mucha<br />
Negative prompt: cartoon, 3d, ugly face, (disfigured), (bad art), (deformed), (poorly drawn), (extra limbs), strange colours, blurry, boring, sketch, lacklustre, repetitive, cropped, hands<br />
Steps: 60, Sampler: DDIM, CFG scale: 12, Seed: 1940895508, Face restoration: GFPGAN, Size: 512x512, Model hash: 7460a6fa, Batch size: 6, Batch pos: 0

Stable Diffusion 1.4

gandalf, d & d, fantasy, intricate, elegant, highly detailed, digital painting, artstation, concept art, matte, sharp focus, illustration, hearthstone, art by artgerm and greg rutkowski and alphonse mucha<br />
Negative prompt: cartoon, 3d, ugly face, (disfigured), (bad art), (deformed), (poorly drawn), (extra limbs), strange colours, blurry, boring, sketch, lacklustre, repetitive, cropped, hands<br />
Steps: 60, Sampler: DDIM, CFG scale: 12, Seed: 1940895508, Face restoration: GFPGAN, Size: 512x512, Model hash: 7460a6fa, Batch size: 6, Batch pos: 0

Stable Diffusion 2.0

Prompt: Portrait digital art of Bill Murray from Scrooged (Arcane). wearing a suit, Christmas,

Negative prompt: cartoon, 3d, ugly face, (disfigured), (bad art), (deformed), (poorly drawn), (extra limbs), strange colours, blurry, boring, sketch, lacklustre, repetitive, cropped, hands

Steps: 40, Sampler: DDIM, CFG scale: 13, Seed: 3408805356, Face restoration: GFPGAN, Size: 512×512, Model hash: 7460a6fa, Batch size: 6, Batch pos: 0

Portrait digital art of Bill Murray from Scrooged (Arcane). wearing a suit, Christmas,

Stable Diffusion 1.4

Portrait digital art of Bill Murray from Scrooged (Arcane). wearing a suit, Christmas,<br />
Negative prompt: cartoon, 3d, ugly face, (disfigured), (bad art), (deformed), (poorly drawn), (extra limbs), strange colours, blurry, boring, sketch, lacklustre, repetitive, cropped, hands<br />
Steps: 40, Sampler: DDIM, CFG scale: 13, Seed: 3408805360, Face restoration: GFPGAN, Size: 512x512, Model hash: 09dd2ae4, Batch size: 6, Batch pos: 4

Stable Diffusion 2.0

What’s next

I’ve started writing up how the Stable-Diffusion WebUI works and will probably wite up how to scrape a google SVG slideshow.  

116

What methods of teaching will be employed?

There are various themes in pedagogy that cater to different types of learner so classes should do their best to reflect these types in the classroom.

There should be a division between lectures and seminars/lab work, or theory and practice/debate. This means information is distributed en mass initally and the discussed in smaller groups for clarification.

As stated earlier (e8) students who are proficient in a topic, mitosis for example, should be able to help assist teaching those of a lower level covering that subject. By explaining a topic one comes to understand it better.

Each high level student will have a mentor he can refer or defer to in times of need and have dialogues with. A mentor must be at least a number of grades above the student.

E-learning will be progressively worked on as will videos created to explain all topics in the subject, complete with questions to assess comprehension and open questions that students might consider writing experiments for, contemplate on or engage in dialogue about.

New methodologies of teaching should constantly be assessed and debated. Also all new members should have some exposure to teaching seminars and lectures [not just participating].

[Note: In the years since I’ve written this the rise of MOOC’s and available resources for learning everything from code to languages has exploded across the internet, in this case these give the community nearly all the material required to created an extremely nuanced hierarchy of grades without lifting a finger. However, all free open projects should be given as much support as possible via the community to continually help and improve their services. I know these initial thoughts were written on buses and trains around town, each one of them could be the basis of whole books in themselves, however the good news is there is a huge wellspring of available resources now to help any community member learn an incredible amount even if they are the first in the community to express an interest in the field.]

 

113

What fields of science, technology and education will the community endorse?

Due to primarily working on a practical development, skills and reasoning platform, there will initially be a focus on the areas of science, research and engineering (computer/electrical). With a broad base of the basic sciences mathematics and logical reasoning. Ultimately, the community should endorse, teach and promote all the legitimate forms of science and even assess the claims of things considered pseudo-sciences as well.

Initially the organisation will be small and this so this may result in certain certain subject or entire departments having only one person. In time these will fill to reasonable sizes and have more members to spread the duties among, the more time can be dedicated to unique and specialist areas.

There should be a special interest in promoting programming and computer science skills to every member as the ability to develop software is useful in almost any scientific discipline.

[note: computer science skills should also be taught as they should be the bread and butter skill of the community, allowing the community to work en mass on projects to improve them imagine having 5000 community members working on improving a project like Ubuntu or LibreOffice and the amazing strides those projects could make with simple bug fixes being implemented.

112

What types of research will the community endorse?

As long as the research does not go against the ethical oath [e57] of the community it should be allowed. Initially there will be limitations of scale and time constraints, represented disciplines e.t.c. But personal research should be given as much autonomy as possible. When it comes to resource usage, available equipment or problems of costs, there must be a meeting with the head of the relevant departments to request a loan, grant or new equipment. A project requirements document is needed [e62]. Conversely while projects can be endorsed that tackle similar subjects there is a focus on collaboration, so having 3 people doing the same thing seems like a waste. There should be a clear database of projects that are currently in progress to avoid such problems (or at least a list of projects internal to the country not including others).

[note: Massive projects and subjects require hundreds of concurrent teams collaborating, projects can also be approached from a number of novel angles. We could take cancer cure research or A.I as two examples where there are a vast variety of methods being applied to try and discover the best way to solve the problem. Similarly pure research into subjects such as climate change have thousands of teams researching a whole host of elements, while two teams doing precisely the same thing is not useful (unless of course there are multiple teams working on replicating research for peer review) there are a number of reasons that multiple avenues of research may run in parallel.

The database of current projects though is a great idea, more than this a list of potential project or research ideas could be added to the list allowing people to browse research that is considered open for review or investigation.]

les sites educatif

aujourd’hui j’ai fait beaucoup. J’ai utilizé les sites-web duolingo et aussi trésbienFrench pendant 3 heures. Ils étaient un peu durs, néanmoins j’ai persévéré, et dans le processus que j’ai vu  il y a un long chemin à parcourir!

En fait, c’est pas tout mauvais, j’ai vu beaucoup l’amélioration aussi; Je suis sûre; si je continue finalment je gagne.