UPDATE - Please check out my LoRA version of this, which performs better on average and is more flexible as it's a 1.5 model!

My first model I've decided to release. I made this out of respect for one of my favorite shows of all time, Star Trek: Deep Space Nine.

Please do not repost or share elsewhere without my permission. If you want to do merges or whatever, feel free, just please don't publicly share without asking or sell them. Also, please don't do anything creepy or gross with this model, and keep the NSFW to yourself (I haven't trained or tried this model that way).

Please DO share your results!

04/29/23 - A few people have posted about issues with this model. Just to note, you MUST generate images at 768px on one edge as I note below. You should also have the 2.0 .yaml file alongside this model file.

This was trained on the remastered shots from What We Left Behind, in January 2023 using TheLastBen's Fast-Dreambooth. Since then, Dreambooth training seems "broken" and I haven't been able to continue training this at the moment. If I find a fix, I plan to further train this on characters and species that could use refinement.

You really need to be a prompt engineer to push this model to its limits. All of my image results are not inpainted, did not use Controlnet/etc., and were not edited in any way. Layering in those will help you get even better results. Please look at my shared results, which include all of the prompts and negative prompts for a wide range of applications.

Here is my basic prompt. This is based on how I captioned the training data (starred are optional):

• sdn, a [closeup/medium/wide] view of [character or object], [species], [gender*], [expression*], wearing [Bajoran/Starfleet/Cardassian] [Operations/Science/Command/Security] uniform*, [lighting descriptors (diffuse glow/contrast lighting/etc.]*, [inside/outside], [location (operations/planet surface]*, [background (blurry/viewport/panel/etc.]*

Use “sdn” as the initial keyword. Not technically necessary, and sometimes overweighs training appearance, but keeps coherence of concepts and characters better. 

768 x 768 is by far better than other aspect ratios or sizes. I recommend doing all generations at 768 and then outpainting. It works fine with wider dimensions, such as 1024x768 or 768x1024 but the short edge should always be 768. Sorry if your card can’t handle it but I figured if I was going to train, I might as well have trained it on 768.

Use terms like “star trek” “ds9” or “screenshot” etc. to further push certain concepts.

CFG of 7 is good for general purpose. Higher for concepts that vary away from original references. Sometimes, things look a bit too intense and contrasty. Take CFG to 5-6 to get a bit of the original film glow back.

Sampling steps with Euler a, between 20-30 look good. I stick to 24/25 unless I want a bit more contrasty crisp look.

Negative prompts help push results towards more aesthetically pleasing generations on average. My personal go-to is:

• bad anatomy, bad proportions, blurry, cloned face, deformed, disfigured, duplicate, extra arms, extra fingers, extra limbs, extra legs, fused fingers, gross proportions, long neck, malformed limbs, missing arms, missing legs, mutated hands, mutation, mutilated, morbid, out of frame, poorly drawn hands, poorly drawn face, too many fingers, ugly

Of course, sometimes you do want deformed or ugly results, so adjust as you need. “Blurry” reduces natural fuzziness of original training so also optional (I negative prompt “blurry” and positive prompt “diffuse glow” for example, to sharpen and keep effect)

Uniforms, specifically Starfleet uniforms, should be keyworded, such as “Bajoran Security Uniform” and “Starfleet Command/Science/Operations Uniform”. Starfleet uniforms don’t always come out in the correct colorways, but can be inpainted if necessary. I also recommend using slight prompting to push what you want, adding in “red black and grey” etc. for example. For starfleet uniforms, Worf’s sash will occasionally appear without prompting. Negative prompt “sash” may help.

Comm badges and pips can sometimes do best with inpainting. Comm badges may appear in duplicates.

My initial training overweighs Bajoran nose ridges in all species, but especially humans. Inpainting usually works well to get rid of them if necessary. Further training I did toned them down but they might be less weighted if I train even further on other species. Negative prompt “bajoran” can help.

Including actor names can really push characters, especially Miles, Ezri, Nog, Leeta, etc. and I recommend it for basically everyone but Avery Brooks since his training worked so well.

Adding “beautiful” and “handsome” etc. can make images look better, I recommend especially for Kira. “young” “30s” and more can push toward what you want.

Bajorans

• Kira looks older for some reason. Use “youthful” “young” “30s” or “beautiful” in prompts, or “aged” “old” “50s” etc. in negative prompts.

• Using “vedic” does somewhat produce the look, but it also skews generations toward real-life concepts of vedics. Using “Indian” as a negative prompt can help.

• Even though I trained and re-trained on Kai Winn, she doesn’t really come out in generations.

• Most Bajorans end up in Kira’s uniform. The security uniform sometimes needs “tan” or “beige” to pop out.

Cardassians

• The model is trained on Garak, Gul Dukat, and other Cardassians but doesn’t seem to want to “separate” them and most of the results resemble Dukat. (I might further train on Garak and others)

• I included some images of Dukat as a Pah-wraith in initial training and the model overweighs red eyes for Cardassians. If you want to, use “red eyes” as a negative prompt and it usually eliminates them but does alter generations slightly from the same seeds.

Changelings

• Even though my training had founders, it’s kind of Odo or no one else. Further training on the other founders would possibly help.

• Odo ends up looking scruffy and rough on occasion, sometimes with lots of ridges, and sometimes extremely smooth and blurry. It’ll take a lot of generations to get him to look correct but as I show, it can be done. Definitely add "rene auberjonois" to your prompts.

Ferengis

• Like Cardassians, their eyes tend to look demonic. Undo it with “glowing eyes” or “evil” in negative prompts.

• It really doesn’t understand the difference between them, even though it was trained on Quark, Rom, Nog, Ishka, etc. Actors' names are necessary. Nog comes out best.

• Use “uniform” in negative prompts for stronger “Ferengi” clothing.

Humans

• Sisko is by far the best trained subject.

• O’Brien needs a bit of extra tweaking, add in Colm Meaney and “full face” for a proper look.

• Keiko doesn’t really come out even though she was in training data. I’ll further train her if I do train more.

• Jake Sisko and Kasidy Yates was in the training data but only Kasidy works with the addition of “Penny Johnson Jerald”.

Jem’Hadar

• Again, they were trained but definitely not quite enough. They only sort-of resemble them unfortunately.

Klingons

• Worf comes out great. Others, not so much. If I further train, I’ll expand my data to include a wider range of Klingons.

• Use “sash” to get Worf’s sash to generate. It can also be generated/inpainted on other characters.

• Worf’s forehead goes a bit wacky but inpaints easily. His nose also randomly turns red. I have no idea why.

Trill

• It’s hard to get the spots to pop up. Inpainting is actually difficult with this too, but trying ((leopard spots on skin)) for example can help.

• Jadzia and Ezri are both in the training but need actor names to really pop. Unfortunately actor names also lower spots. 

Vorta

• I trained on multiple Vorta, but it just spits out humans with light skin and black hair. More training will be necessary here, unless you can prompt engineer them.

Other Species

• No training on Andorians, Benzites, Betazoids, Bolians, Orians, Vulcans, Romulans, etc. Some would pop out with other embeddings or keywords that the general training data knows.