🏞｜general-with-images

[boreal crown]

if you look for that image with tinEye the amount of photoshopped touches to it are hilarious.

[bitter spruce]

I think today i'll take my fav wallpapers and try some img2img

[spice parrot]

how can I generate images here?

[bitter spruce]

here? like... right here?

idk

can you?

i don't see an image create button

what does your screen look like

[spice parrot]

Generally, how can I generate images using Stable Diffusion?

I am asking because I generate images on MidJourney in the chat

[bitter spruce]

hmmmm

your question is still invalid

if you've used a web-ui before i don't understand how you don't know what to do

it's the same as it ever was

[spice parrot]

https://beta.dreamstudio.ai/dream

Got it

[bitter spruce]

ok.....?

weird question

"I'v done this before, but I totally forgot"... "How do I do it?"

lol

is how I took it

https://tenor.com/view/headache-migraine-stressed-out-charlie-day-its-always-sunny-in-philadelphia-gif-5048553

[spice parrot]

Thank you

[bitter spruce]

yw?

https://tenor.com/view/heyar-heyartv-merci-beaucoup-gif-21433673

pc soft crashed

went bk to my login screen

😒

this happens to me when I switch models sometimes

kind of annoying

at least its not a hard crash

hmmmm, i think something is fucked up

😒

wow what a great model 🙄

4 different models doing this now

I think txt2img is broken on my side

going back to img2img

elldrethsLucidMix_v10

still garbage

anybody know a good Armour model?

like knight amour and the like

and what is "loop back" mode with img2img?

[static tusk]

loop back will take your output image and put it back in as the source, in a loop

so you can iterate changes

[bitter spruce]

ooooh ok, so if you set it to 3 images, each final image is the start image for the next, correct?

[static tusk]

yup

[bitter spruce]

cool, thnx for clearing that up

job for a cowboy? nice

i bet a lot of ppl can't rattle off where you got that pfp from huh?

[bitter spruce]

https://youtu.be/lZUMTINGVuo

using same LucidMix model

might try the vampire model, it looks ... so/so

[bitter spruce]

looking pretty cool ngl

[fresh hound]

everyone needs a drink after work

or a vodka!

[wispy nest]

hit or miss?

[bitter spruce]

the baby hand, 😂 i'm fkn dying

That's great lol idk y i think that's so funny, XD

[fresh hound]

I know right hahaha

there is something about the way the camera looks when I use kermit that I really really like

[bitter spruce]

making a server for ppl with art more like me, with a website landing page

😁

even made a logo lol

Prompt Studios, S.D. (18+) SD=Stable Diffusion

I think it's pretty sick so far

I'm also trying to rebuild invokeai's web-ui with SD 2.0 instead of 1.5

muFuggin 1 man devops, webops, design, and everything else. lol 🙃

Prompt Studios SD18 😏 that shit even sounds cool

( If you can't find what you want, make it) I guess is the moral of the story.

The second circle icon 😂 better not

it's not AI art but it is art. Logos n shit. I dunno, I don't want to pervert the art space you all created here, So I'm doing my own thing, just shring 'cause... I dunno, I don't have anywhere else to share it. ¯_(ツ)_/¯

[bitter spruce]

😅 sick the site should be fully up within 24hrs

still need to finish the discord server

preview link http://my.sites.logo.com/preview/997f645b

it looks dumb on a tablet or phone, looks best when viewing the monitor mode

prompt studios 🤣 so dumb

might as well take the meme and run with it right

😒 waiting sucks but the domain is legit and finished woohoo

www.promptstudios-sd18.com XD

That's how you get shit done frenz

instead of me continuing to bitch about other servers nsfw content being not allowed

[gloomy elbow]

You're a problem solver, I like that.

[bitter spruce]

ohupdate to me adding sd v2.0 so far so good

[bitter spruce]

thnx bro

if i'm gonna do something, go all the way

logo, website, server... fuck it, get it done ya dig

anybody here good at discord server disign and wanna hop on this while its fresh lettme know

there's a lack of adult AI places to go online, besides "unstable-diffusion's" server

[gloomy elbow]

on the topic of solving problems, Here's a reminder that if anyone wants an AI training dataset based entirely on open content, https://imagemonkey.io is a crowdsource attempt to provide that dataset, if you are complaining about copyright, you could be using that energy to upload, label, or annotate a solution.

[bitter spruce]

biff bam boom, a man with the plan

nice

< Hi fives @gloomy elbow >

awww I thought the site was up, it just loaded then failed, it's getting close

like waiting for something to get out of the oven

but yea, fellow AI artists, open enrollment for an adult AI art server. I could use some ppl, And i'm not trying to be rude or disrespectful to this server at all. Infact the opposite.

I'm trying to be 100% respectful

or if you got that ... 1 friend artist you think would be with it, tell'em

This AI art stuff came out of fucking left field and I apreciate this place being one of the first

it's kinda the wild west for AI art right now

i'm just setting up my AI Art ...hmmm Bordello? lol

https://tenor.com/view/skofqq-gif-26055170

[mental pollen]

dreamt of zoo

[alpine gust]

Hello all, where u created this pictures?

[bitter spruce]

I got all of SD2.0+ working in invokeai's webui
just waiting for sdv2.0+ weights to finish downloading

🤓

granted a little messy it should look like this on the surface minus a few folders like stuff obv

i make a fake pyproject.toml just to bypass warnings

well, not fake... just empty

move all the v1 configs out and put the v2 configs in

more or less, just do that, with the sdv2.0+ files on github, and update your conda env, install the pip stuff and download v2.0+ weights

easy

oh ya, xformers are also fully working

😉

on AMD

link to weights https://huggingface.co/stabilityai/stable-diffusion-2-1

link to sd2.0+ https://github.com/Stability-AI/stablediffusion 1/2 way down the page

[alpine gust]

Thank you!

[bitter spruce]

and run everything under Requirements

ls if the new git clone has an ldm folder in it, just copy all of it to the original ldm folder

do that with everything, swap it.

then run conda env update/ect ect, then requirements

now you can swap out whatever new SD they come out with

just follow that same bunch of nonesense i just said lol

(fucking scatter brain)

obv i skipped the cuda part cus i'm amd

like pay attention to this stuff

if you use what the project has, it uses sdv1.5

same with invokeai's webui

just swap it out

Key note, I make this up as I go, I know enough to be comforatable editing in any prog lang, but i can't like really code in any lang either

as long as you can make a simple loop in shell, you can figure your way around any lang

all langs are really just same shit

[gloomy elbow]

messing around with portrait prompts and different cfg scales. at some point the nose just disappears!

[pseudo aspen]

Can you help me identify what model was used to generate this picture?

[dry crow]

@pseudo aspen i tought you created it with anythingv3?

[pseudo aspen]

This is not image generated by me. I am trying to reach this level of details

[dry crow]

Ahh

Okay

[pseudo aspen]

But even though I tried many prompts (and the images look very good though) I am unable to get reflections etc. like above

[still flax]

did you try the png info tab?

[pseudo aspen]

No. I'll try that

[dry crow]

Not working

It was edited

[still flax]

maybe discord gets rid of the metadata

[dry crow]

Nope

Can Download other pictures and look at the meta data just fine

[thorny zephyr]

Ah so you can see where people take the pics they send

[dry crow]

It gets overwritten when you use other photo editing tools or even cleared if you use an exif cleaner

[still flax]

just if they are using a phone with location embedding active

[dry crow]

Yea or the prompt and settings

[thorny zephyr]

Most people do i think

[pseudo aspen]

I will try to download it directly from Twitter

[still flax]

I think twitter does remove metadata... But still you can try

[dry crow]

@pseudo aspen getting closer

[pseudo aspen]

Whoa. Yeah

[dry crow]

@pseudo aspen very close:

with no img2img

[pseudo aspen]

Wait

Really?

[dry crow]

yea i just tried xD

some tags that i think would be used

[pseudo aspen]

Could u send me an example of ur prompt?

[dry crow]

the first image is on a model merge of mine, the second one is done with AbyssOrangeMix2_nsfw xD

[pseudo aspen]

Hmm interesting

I'll get this one

[dry crow]

yea i think that could have been used for that image

[pseudo aspen]

Yeah. I think you found it :D

[dry crow]

here are the prompts without the quality tags:
1girl, teenager, (white marine admiral uniform:1.3), gloves, black hair, long hair, straight hair, black fox ears, (:D), (port and city in the background:1.3), reflection, waves, wood planks, (night:1.3), big breasts, city lights, cowboyshot,

[pseudo aspen]

Thx a lot

[dry crow]

np! 😄 send me some good ones, going to bed now

[pseudo aspen]

Same! I will send u

[wispy nest]

How can i solve this?

[lofty flax]

Hey guys, how can I make a anime girl from a logo or meme like the famous troll guy anime girl?

[hasty nova]

I got that same error this morning, must be an issue with auto1111

or a certain extension

What extensions do you have installed?

[dry crow]

You can put it into img2img

[lofty flax]

where necessarily? I'm new at this

[dry crow]

Are you using SD local or in other Form?

[lofty flax]

online via studio

I have GTX1650 mobile so I can't say if I can use local

[dry crow]

With 4gb vram? Then you can use it local

[dry crow]

If there is an img2img Option available in this studio then try that. I think Collabs would have that.

[lofty flax]

I tried but doesn't give best results

[lofty flax]

I can? So I'll try, I think i need 8gb vram or something

[dry crow]

No with 4gb it will work, set it up for a friend with 3gb 1060 xD

@lofty flax for a good Tutorial Start here:
https://m.youtube.com/watch?v=VXEyhM3Djqg

[lofty flax]

very niceeee

thanks bro

[dry crow]

Np, when he is talking in the Video about Adding --xformers you will have to add --medvram also (seperated with a space)

[gloomy elbow]

I love how the tail in this concept communicates that it's not an anthropomorphic cat, it's a regular cat controlling a robotic body,

[flat kernel]

They deleted the boobies and with them most of the good anatomy

[gloomy elbow]

Takes a ton of getting used to, prompts feel totally different to me

[dry cosmos]

My images are getting grayed out, does anyone know what it could be?

[dry crow]

Youre missing the vae file for that model

[sullen blaze]

A comparison of 43 prompts in midjourney and dalle2 that can help you master yours and inspire new styles. https://saipien.org/dalle2-vs-midjourney-art-prompts/

[boreal crown]

this is cool but you prompt dalle, midjourney in different ways to achieve results. its kind of like saying dogs are hands down better than cats at fetch 😉

im a sucker for those little peel away reveal things on webpages though

[bitter spruce]

don't forget about adding and removing weight to your prompts

(positive weight)+
[negative weight]+

(positive weight with less importance)-

[negative weight with less importance]-

and mixing

example: {cat|dog}, [outside], (looking out the window)++, {birds|crows|[people]--}

you can play with that, see what it does, add or remove words and note the output, try like {day|night}

I forget what SD is calling this method of prompting.

like... if / or

{if|or}

left curly bracket, prompt word, pipe, prompt word, right curly bracket

[south gorge]

Can you give an example

[bitter spruce]

That site I made, still not up, I had to email them and let them know that if it is because of incompetence, and the site isn't up in the next 10 hours I want a refund. ugh, never fun making one of them emails 😒

[bitter spruce]

I did.

do you see the word EXAMPLE ? (in lower case.)

all of that text above your comment IS the EXAMPLE(S)

[bitter spruce]

and I'm being nice, lol

I could have easily just created a subdomain for free from one of my own many websites ya know. but I choose a specific domain name, and If the site builder wants to automate the process fine, but make sure your shit fucking works, lol, for fucks sake. 🤦

[pseudo aspen]

How do you guys merge models

[bitter spruce]

by training

or lookup that key word on stable diffusion's github wiki page

I was able to make a few checkpoints using auto1111 but invokeai you have to do that manualy in the terminal via the provided scripts

look for README.md files in your webui's folders, and read them

they are in all caps and named READ ME on purpose. For you to notice them and read them.

@pseudo aspen If you cannot figure it out after my advise to help yourself, I will find the answer for you.

but you should also search for that by using this servers search on the top right.

[pseudo aspen]

Thanks a lot!

[bitter spruce]

yw, I have faith in you.

idk how much help i'll be in a min though, I went to the cannabis shop today and spent like $200 on topshelf nom noms, lol

[pseudo aspen]

lol

[bitter spruce]

https://media.discordapp.net/attachments/1048866961670549566/1063706790719660052/drug-anime.gif

AI Art is cool

[pseudo aspen]

Indeed

[bitter spruce]

I could try to find what I'v made that is sfw to share

1 min, i got these oil, hash, rolled in kiefe balls to add to this nuggy bowl

lol, got an email back, "we assure you, bla bla bla"

just fucking do it, don't tell me you are, lol dip shits

omg, I think I forgot how to mask for inpainting

the in paint adds and the mask preserves right?

[pseudo aspen]

Dude why is stable diffusion always generating something lewd for me

[bitter spruce]

or am I backwards?

[mental wolf]

Put together a little demo of how to visualize the internals of the depth2image SD 2.0 pipeline: https://twitter.com/NikolausWest/status/1616013591051739136?s=20&t=TuLJ-x3wurel99sgKwN-8g

Nikolaus West (@NikolausWest)

Building @rerundotio has made it painfully obvious how much time I've wasted trying to introspect my vision code in the past.

Here's are the couple of lines needed to look under the hood of the depth2image #stablediffusion pipeline from @huggingface

▶ Play video

Twitter · Jan 19, 2023, 10:03 AM

[bitter spruce]

what models are you using?

oh, mask preserves

[bitter spruce]

where are you getting your models from?

there are nice models here https://civitai.com/ but .... be warned now, there is mostly Lewd stuff, because the worlds nerds, don't get laid enough, and the first things being made is to fulfill ones fantasies, before functionality

[pseudo aspen]

Thx

[bitter spruce]

lol, yw

and because, most programmers are men, its in our nature to be lewd

lol

facts ¯_(ツ)_/¯

rewind to the 90's when dudes were making millions overnight by creating the beggining of internet porn

[pseudo aspen]

Like I didn't say I mind lol. But I don't want it every time

[bitter spruce]

OOOooohhhh

i can help you

i have some word lists and you can use them to create your own negative word list, and just put [this] around the word

so, example: [nude]+++

this will tell SD that NUDE should be 3 times taken away from the final image

(beautiful dress)+++ this ='s Tell AI i realy, realy, realy want a beautiful dress in my final image

realy x3

lol

[pseudo aspen]

(perfect hands)+++++++++++++++

😄

[bitter spruce]

😉

tell me how that works out for you

[pseudo aspen]

Yeah. Thanks

I appreciate ur help

[bitter spruce]

https://tenor.com/view/neko-dance-maid-gif-11478139

yw

(that's just my happy gif)

😒 i suck at inpainting all of a sudden

took me a min to find something i can share

i like tattoos

https://civitai.com/models/1136/postapocalypse

this one is cool looking

NEW and i'm totally getting it GalaxyTimeMachine's GTM_UltimateBlend_v2

[versed furnace]

Been working on one as well. Sort of mad max meets fashion week.

[bitter spruce]

that's cool as fuck dude, nice

[versed furnace]

2.x though

[bitter spruce]

idk wut that means

[versed furnace]

The post apocalypse checkpoint you linked is stable diffusion 1.5. I've been working with 2.1

[bitter spruce]

ooohh

yea me too

you can use 1.5 with 2.0+

i don't see much of a fdifference

[versed furnace]

none of my embeddings will work with 1.5 models

[bitter spruce]

really?

add in the launch command

I use invokeai not auto1111

[versed furnace]

Invoke doesn't support 2.x at all last I checked.

I haven't looked in a while, thogh

[bitter spruce]

--embedding --embedding-dir=/path/to/embeddings

[versed furnace]

Have they updated it?

I'll need to go check out that project again

[bitter spruce]

yeah, it works well

cool, SD is coloring in tattoos to match hair color, that's intersting, pink hair is also creating really nice pink tattoos

(that would cost so much irl)

uploading new creations on my gallery (follow link by clicking my profile picture, get a better idea of what I mean about pink hair + combined with tattoo prompt) WARNING my stuff is for adults so Viewer Discretion Is Advised

I really like the evil snow white with a border

tho, my art is tame compared to other stuff being made

oooh i know, i'm gonna start to morph the lastest image i made into Zero-Two

it was a no brainer

[bitter spruce]

nice

Tip for invokeai, open model manager click on a model and add the vae.ckpt to the model and update then models that failed to load will load

you can also change which config the model uses

v2-inference.yaml instead of v1-inference.yaml from the configs folder

or the inpainting.yaml config

[azure pine]

horse

[lofty flax]

How I can enhace the characters face?

I only have 4GB vRAM

[pseudo aspen]

I have 2GB and intel HD 4000

[lofty flax]

Just... How?

[pseudo aspen]

I don't know... It just... works

Somehow

[lofty flax]

Which models?

[pseudo aspen]

Anything-v3/4/4.5

[lofty flax]

Gonna try

[lofty flax]

SD web ui?

[pseudo aspen]

Yes

[lofty flax]

Another question 😅, using anaconda or git and python?

[pseudo aspen]

Git and Python lol

[dry crow]

the character has to be more in the foreground, to get more details. Also you can use face restore, with gfpgan

[lofty flax]

gotcha

I got this, but only after set the denoise strength to 1.0

[lofty flax]

I'll try to focus in foreground, the gfpgan seems to be more complicated to me

[frigid vigil]

@gloomy elbow wait is there a different prompt structure? Is there ANY advantage to 2.1? If not im just gonna stick to 1.5

[waxen remnant]

Hi guys, what is the best method for style transfer?

[gloomy elbow]

Not structure so much as the way weights are evaluated are differently. I make sure I use it a little each day since that method of evaluation will probably be the basis for future development

[frigid vigil]

@gloomy elbow but is there any advantage to 2.1 over 1.5?

[dry crow]

yea there is an 768x768 model only for 2.1

also depth model

[frigid vigil]

@dry crow depth model? Explain plzzzzzzzz thanks

[dry crow]

cant explain it better than here: this goes for 2.0 and 2.1
https://stability.ai/blog/stable-diffusion-v2-release

there you have all new features of 2.x versions

[valid stratus]

🌟hi there..just joined,where and how to start?😀thanks

[wispy nest]

I'm going to try my hand at making some custom vinyl wraps for some guitar repairs I'm doin, if you guys come across any neat textures or patterns send em my way!

[dry crow]

Hey welcome, depends on what you want xD

[dry crow]

@drowsy plover Youre in the Stable Diffusion Discord and not in the openai dc.

[drowsy plover]

oh

yeah real

i forgor

nvm

[valid stratus]

Create some Images

Other places have only limited rooms for newbies

[still flax]

#1014939219904450590

[junior gust]

It's crazy how good stable diffusion is getting with all the model merges floating around

I've been trying a ton more models since I got on civitai

[frigid vigil]

@junior gust civitai?

[junior gust]

Yeah check it out, great place to get models and textual inversion files https://civitai.com/

[frigid vigil]

@junior gust ah hell thanks dude. Next time reply at us directly tho it helps us backtrack in case we miss a response. HELL YEAH THANKS

[junior gust]

Of course. np

[frigid vigil]

The regular library has barely anything. Only cool one i found was silent hills Pyramid Head. This should be great.

[gloomy elbow]

Man, gym goals:

[spring cape]

having a weird issue generating images, anyone see the last step suddenly go burnt? https://imgur.com/a/getpKrZ

[dry crow]

so hyped about this:

[rancid pulsar]

blend

blend

imagine/• alchemy, ripley scroll,mutus liber, atalanta fugiens, voynich,laboratory, athanor,books, alchemist, Nicolas Flanel

imagine/• alchemy, ripley scroll,mutus liber, atalanta fugiens, voynich,laboratory, athanor,books, alchemist, Nicolas Flanel

[junior gust]

I've had that happen with certain model and sampler combos. Switch sampling methods and the issue should go away.

[wispy nest]

where is inpaint full resolution

[dense haven]

Anyone know of a decent model that generates proper images of piano playing? Mine's having a hard time

[wispy nest]

has anyone tried using 3d generated wireframe models of rooms as an input to images?

i'm getting funky results that look really AI generated:

(depth2img seems to give slightly better results than img2img)

saw this thread and tried to give it a shot: https://twitter.com/justLV/status/1605276764598665217?s=20

Justin Alvey (@justLV)

I used the #StableDiffusion 2 Depth Guided model to create architecture photos from dollhouse furniture.

By using a depth-map you can create images with incredible spatial consistency without using any of the original RGB image.

See 🧵

Likes

606

Retweets

127

▶ Play video

Twitter · Dec 20, 2022, 6:59 PM

[tranquil kindle]

@ornate wagon same here

[open sun]

@gloomy elbow there there

rogue is sigma AF

[stark scroll]

2.1 just wins when inpainting faces

[pallid relic]

civit seems back up

[ornate wagon]

Looks really nice. Is that vanilla 2.1 or a merged model?

[ornate wagon]

Also, love the style. Could I ask for the style prompt?

[wispy nest]

the stuff they used can be found in the image's meta data :)

[ornate wagon]

I didn't get anything when I checked the parameters, but I'm probably doing it wrong. 🙂

[dry crow]

@wooden dagger here we go

[wispy nest]

did you download the full image? or the preview in discord?

and if it doesn't work, just open the image using notepad, then you'll see everything they used at the top

[wooden dagger]

I get goosebumps

[dry crow]

😄 compared the most used samplers, but i think there isnt an advantage for 60 steps, if i upscale

[wooden dagger]

@dry crow do you mind sharing the prompts?

[dry crow]

its nothing special, here are the tags but without quality tags:
1girl, teenager, (contemporary clothes), styled, cool, beautiful, makeup, wristwatch, chain, earrings, Jewellery, trendsetter, outlet city, branded, cowbowshot,

[wooden dagger]

do you have the seed also?

[dry crow]

88888888888888888 xD

cfg 7, steps 30

or 60

[wispy nest]

did you test the 88888888888888887 before it? :P

[dry crow]

no 😮 but i can, but now only 30 steps, faster for me

[wooden dagger]

that with -1 seed

[dry crow]

on sd2.1 ?

[wooden dagger]

yes

but again I did a beginner mistake

[dry crow]

i didnt describe the eye color so its funny that she closes them for the makeup tag xD

[dry crow]

which one?

[wooden dagger]

I did use the deafult model

[dry crow]

yea 😄 much better now

[wooden dagger]

but but...i want anime

[dry crow]

then you need an anime model

can recommend, AnythingV3, ACertainModel, WaifuDiffusion 1.4

[wooden dagger]

oh ofcourse, do i need to say it, I am STUPIDO

[wooden dagger]

where do I find those , are they on CIVITAI ?

@fresh hound not bad

[dry crow]

@wispy nest wow i typed 88888888888888887 as seed but it did 88888888888888880, (mouse scroll trolled me)

[wooden dagger]

hmm.. that was strange

[dry crow]

very strange

i can link you the models

[wooden dagger]

I found them

downloading as we speak

[dry crow]

for ACM: (ACertainModel.ckpt)
For Anything download: (Anything-V3.0-pruned-fp32.ckpt

• Anything-V3.0.vae.pt)

[wooden dagger]

but where did you see that it used the wrong seed?

[dry crow]

under the generated images there will be a full output of all settings used

ah i found the cause of the bug xD, i was in the seed lane as i wanted to scroll the website up, scrolling will change the input numbers

[dry crow]

make sure you have the .vae.pt file in the same folder as the ckpt. then rename them to match the name

like this:

then the image colors will be a lot better

[wooden dagger]

ah ok

those I cant find

[dry crow]

https://huggingface.co/Linaqruf/anything-v3.0/tree/main (Anything-V3.0.vae.pt)

[wooden dagger]

thank you

[wispy nest]

i want fucking artstyle animation style anything but every time i get wierd realistic shits

[dry crow]

maybe you want to try samdoesarts models?

[wispy nest]

let me check that

[dry crow]

or you mean anime style?

[wispy nest]

dude i just dont want realistic

like 1.5

[dry crow]

yea then go for anime models or artstyle models like inkpunk

[dry crow]

is it working now with the vae?

[wispy nest]

ok i didnt use stable diffusion from 1.5 in 1.5 getting realistic was hard

[dry crow]

i think 1.5 did well with realistic

[wispy nest]

yeah but you should want that if you didnt type realistic it will give you cool art styles

[dry crow]

yea but it was trained on to much random stuff, not specific artstyles

you could also check Dreamlike v1.0 for a cool style

[wooden dagger]

@fresh houndAquaman as Flash, nice

[wooden dagger]

I quess it is possible also to take a picture of my self, and use that? is that Inpaint or what?

[dry crow]

yea that can be done with inpaint or training

[wooden dagger]

so I could be Flash?

[dry crow]

sure 😄

[wooden dagger]

or Bill Gates LOL

[dry crow]

haha how would someone know that your head on an office guy is bill gates 😛

[wooden dagger]

exactly

only Bill Gates would know

[dry crow]

haha maybe

[wooden dagger]

well there is endless possibilities I quess

[dry crow]

yea exactly, the community is making amazing stuff out of SD

the tool got so advanced

[wooden dagger]

I am both amazed and scared

[dry crow]

im excited, give this 6 months again and we can generate maybe quicker, with even more features and cool stuff

[static tusk]

nvidia said they will release an AI driver improvement soon. hopefully that adds some efficiency

[dry crow]

ouh nice, but i dont think they can help my GTX1080 (missing ai cores)

[wooden dagger]

@static tusk aha nice

I am seriously thinking to buy I new Graphics card just for SD

[static tusk]

I have the rtx3080. does a great job at this sort of stuff

[dry crow]

i go full amd for my next build

nvidia is joking with the prices and their community hate (DLSS 3 only for rtx 40X0) xD

[wooden dagger]

@static tusk RTX is highly recommended

[static tusk]

yeah I think prices are high because NV made a crapload of 30series chips during the mining boom. so they are trying to sell those off first

[dry crow]

exactly, then they also wait to sell the pricier cards first until that they wont release midrange cards

[static tusk]

mid range coming feb-mar

[dry crow]

yea midrange for 900€+ xD

[wooden dagger]

@static tusk Have you tried Canvas?

[static tusk]

I havent

[wooden dagger]

I wanted to be could not because I dont have a RTX

[static tusk]

what does canvas do?

[dry crow]

yea same, seems like a fun tool,

[static tusk]

is that painting?

[wooden dagger]

@static tusk https://www.nvidia.com/en-us/studio/canvas/

[dry crow]

it creates images of your "ms paint" paintings

[static tusk]

ah yeah I saw that. very cool

[wooden dagger]

you could have a child paint a landscape 🙂

[dense tapir]

The problem with AMD is that it is just now getting into ML so RDNA4 is supposed to be all in, but right now training on AMD is pretty dead from what I have read.

[fierce crest]

welp im bored

time to hop on 🟧 ◼️

⭐

[still flax]

I love this kind of aesthetics

[still flax]

Space pirates!

[still osprey]

Via my custom trained model

[stark scroll]

It's vanilla, img2img from a midjourney base

Manga style

[wispy nest]

wtfff I can't generate images no more because not enough vram even though I haven't changed settings

[dry crow]

look on the taskmanager, maybe the vram didnt get cleared, then try restart webui, sometimes it hangs and then the vram get stuck

[wispy nest]

it's used somewhere

I don't understand

[dry crow]

where?

close other applications that might use it

[wispy nest]

is it "memory" in task manager?

[dry crow]

no its dedicated gpu memory

switch to the process tab and click on the gpu header

[wispy nest]

1%

gpu there

[bitter spruce]

you should be using the software that matches your gfx card

[bitter spruce]

to get the best performance and for solving issues

[dry crow]

@wispy nest did you closed the webui cmd?

[wispy nest]

I just did and now it went to 0

xd

I'm very stoned, but I still can't generate images, I restarted pc, I stopped everything

same settings different day

[bitter spruce]

I'm about to get there

>.<

[dry crow]

what error do you get ?

[wispy nest]

not enough vram

[austere pilot]

LOL

[bitter spruce]

error: lighter, not found

[wispy nest]

error I have a lighter and desire to get high brb

[bitter spruce]

unstable are a bunch of sissies

[austere pilot]

unstable has more members then stable lmao

[bitter spruce]

they don't like ppl that stand up for themselves

[austere pilot]

? isn't their purpose to stand for nsfw?

[wispy nest]

Some people just don't like people

[bitter spruce]

yea, and they all liked my art there

and I got into it with someone there that I was helping, and they silenced me, when I could speak again, i told the mod how I felt about there decission and i was banned

[austere pilot]

that's a personal problem

your generalizing the whole community

[bitter spruce]

wtf are u talking about

[bitter spruce]

im not generalizing an entire community u dumb fuck

i called them sissies

[wispy nest]

are you against dumb fucks?

that's dumb fuck'ish

[austere pilot]

lol its time i head out you cant win against an idiot XD

[bitter spruce]

lol

get fucked

[wispy nest]

totally

bro moment

[bitter spruce]

totally

whenever someone speaks against authority, bam ur banned

[wispy nest]

I'm way too happy considering the fact that I can't generate images

[bitter spruce]

narcissistic ppl

u can't?

[wispy nest]

universe hates me, one day I can another I can't

[bitter spruce]

there are sites you can use if you can't get it working locally

[wispy nest]

I guess I'm going back to 1.5 then

[bitter spruce]

https://creator.nightcafe.studio/

what's wrong with 1.5?

and what does 2.0+ have that 1.5 doesn't

[wispy nest]

Oh I didn't really mean it that way 😄

I just wanted to learn the 2.1 768

but suddenly it no longer works

and I'm quite fresh I'll admit

[bitter spruce]

idk wut 2.1768 is

you'll have to enlighten me

or meh whatever

the nightmare never ends

[wispy nest]

US in 1year

[bitter spruce]

?

[wispy nest]

looks like a statue of liberty 😄

[bitter spruce]

i guess

i kinda see it

are you not from USA?

[wispy nest]

No

I'm an European boyo

[bitter spruce]

if not I can see your animosity

ok, that makes sense

[wispy nest]

I wouldn't say animosity

more of a 30T debt

[bitter spruce]

yeah, I would hate on USA too if I didn't have the freedoms USA citizens have

[wispy nest]

which one?

tell me one I don't have 😄

In Europe we call it American ignorance

[bitter spruce]

right

which doesn't sound ignorant at all

lol

[wispy nest]

The freedom thingy is a laughing matter here 😄

[bitter spruce]

sounds more like deep seeded jealousy and hate

and you make stuff up to justify your hate

[wispy nest]

I mean I live in the best per capita country in the world

Im gucci

what I make up?

cmon man

[bitter spruce]

c'mon wut

you clearly aren't up for debate

[wispy nest]

I'm telling you the perspective of European

you don't debate

[bitter spruce]

you just want to fight

[wispy nest]

you didn't give a single freedom that you have and I dont

[bitter spruce]

nothing I say would change your mind

[wispy nest]

and said that I'm jelous xd

[bitter spruce]

so, u just want to fight

i get it

i'm a man too

like bro

[wispy nest]

you implied animosity

I said no

economy

[bitter spruce]

yea, but i also said i was hi

[wispy nest]

same

[bitter spruce]

so, if you want to argue with someone thats hi

lol

[wispy nest]

whatever the fuck that means xdd

We gucci

?

[bitter spruce]

it just means you are right and i'm just fucking around

[wispy nest]

Europe as good as US? 😄

[bitter spruce]

i dont care

jesus everybody online get so upset so easily

i swear

[wispy nest]

Dude you're forcing me to write about projection 😄

[bitter spruce]

i am, then ima jedi

If i can force you to do that, 🤔 what else can I force???

anyways

[wispy nest]

Dunno, I'm very impulse anything can happen

[bitter spruce]

i get that way sometimes

so, how was your day

[wispy nest]

it still is

[bitter spruce]

hows it going so far?

looks like EU weather in that pic lol

[wispy nest]

Can't be that bad xd

[bitter spruce]

that's good

[wispy nest]

There are no hurricanes here cough cough

[bitter spruce]

see, right there, see what you are doing

but if I engage, I'm then the bad guy, right?

go fuck yourself, lol

[wispy nest]

You start it the 2nd time and get mad at my response the 2nd time and then imply that I started the 2nd time

xd

[bitter spruce]

i'm not mad bro

i'm Hi

I'm higher than bird pussy

[wispy nest]

are they your pronouns?

[bitter spruce]

and that shit is pretty high

[wispy nest]

dude

that's actually hi af

[bitter spruce]

mhm

https://media.discordapp.net/attachments/1048866961670549566/1063706790719660052/drug-anime.gif

got sum "spiked punch" at 27.2% thc

[wispy nest]

Oh I bought Critical kush, white widow and candy kush

[bitter spruce]

and my oil kief balls at 44.6% thc

[wispy nest]

don't know the strenghts

my bong is long

[bitter spruce]

nom noms

super dank

weed looks crazy when u zoom into that shit

[wispy nest]

my phose is too shitty for pics xd

looks basic af

[bitter spruce]

womp

i buy my phones from swappa.com it's used phones but everything is high quality second hand, because you have to (by choice) send a picture of the phone after it gets to you, so the seller can keep there rank

it's a pretty good little system they have

i'v never been let down from buying off swappa

idk if they have one in EU or something like it, probably do

like, i'm not gonna spend $1,000 on a new phone, that's stupid

but i will wait a year and buy it for $200-$300

save almost $700

[atomic bough]

But is it trust worthy

[bitter spruce]

fuck yes

yes

i'v bought like 4 phones through them

all in damn near mint quality

[atomic bough]

Do you buy mint

Or good

[bitter spruce]

booth

and booth times they looked new

[atomic bough]

That's good for you ...

[bitter spruce]

you can look for yourself

https://swappa.com

[atomic bough]

I mean

Why would someone sell his new phone if it's ok

[bitter spruce]

it's a part of https://xda-developers.com

[atomic bough]

U can buy old phones with low prices

But new models ...I wouldn't trust them

That's the problem for me

[bitter spruce]

xda is where i find all my custom roms

cause i prefer a rooted phone

[atomic bough]

Yeah

Some phones can't be rooted

No matter what

[bitter spruce]

and on swappa you can ask the seller to put on a custom rom for you

they are cool like that

[atomic bough]

Yeah

[bitter spruce]

one phone i bought the guy put on cyanogen rom for me

i'm not that big into modding roms much these days, but i was really into it durring android 5.0 - 8.0

and once you go root, stock android is just bleeeeeh

you guys ever figure out if civi???.com is still working @atomic bough

playing around with an anime model set

technically, I turned Jeannie from mxrplays into anime

....kinda looks like her but nah

this one is cool I made

looks like a freeze frame from a summoning

even though my art is just art

it's not porn

maybe that's why in society boys think they are girls

who knows

no one have ever told them

or maybe it's because society thinks boobs are digusting

idk

but those who censor art are as just as guilty as nazi's from ww2

o/o/o/o/o/ more censorship mien fuhuer

https://youtu.be/pzDgaGTARqE

[wispy nest]

don't sleep on poop
t.google

[wispy nest]

you're censored to hell and back and the first thing you compare that to...is nazis

[sullen pecan]

Really bad advertising practice, and for that reason BlueWillow, I'm out.

https://tenor.com/view/for-those-reasons-im-out-shark-tank-mark-cuban-gif-16319251

[waxen flume]

@graceful escarp @limber remnant @tropic shell Ban this guy cmon.

[graceful escarp]

On it.

[dry crow]

there are multiple accounts that do this

[graceful escarp]

I have my DMs open for people who need me and I've gotten three of these already.

[tropic shell]

Yeah; we are trying to squash them.

If you see them, please just let us know. Thanks!

[still flax]

thats the cause... Not a good practice tho

response

its not their fault... Ppl just really want the prize

[grizzled sage]

ehhh it is their fault though? since their intention was exactly for people to bring mor users to their server through spam. it's not a new technique to run a contest, and it's long been known as a way to "spam without doing it yourself". The ethics on that kind of community building have long been talked about.

[wooden dagger]

anyone tried Prompt Converter at Hugging Face ?

[fresh hound]

@limber remnant @tropic shell just a heads up ...

[tropic shell]

ty!

[solemn flare]

restoring faces ain't easy, those rock pics are $

[dry blaze]

First proper generation on a steam deck

[hasty nova]

My latest attempt at an embedding is almost done

well, I say it's almost done. I set it for 5000 steps but imma stop it at 4000 I think

[hasty nova]

Little sample from my embedding

[hasty nova]

I released it! #1066559052328472596 message

Super proud of how it came out

[tawny forge]

how can I fix this?

the car not in the image properly

[hasty nova]

negative prompting?

[tawny forge]

how would it looks like?

[hasty nova]

something like "out of frame" in the neg prompt might work

[tawny forge]

alright

dont seems to work

[hasty nova]

hmm, it's really a lot of trial and error.

And some other negative prompts

There are prompting guides on the internet

[hasty nova]

I'm really proud of mt new embedding, it can make some really interesting portraits

Supposed to be a ghost girl

Can also do some spooky stuff

[gritty cedar]

wow thats awesome

[hasty nova]

Thanks! There is a download if you are interested

#1066559052328472596 message

[gritty cedar]

make it a wider aspect ratio

like 850x500

[tawny forge]

I cant handle that

[hasty nova]

A comparison between the two embeds I am offering

Another comparison

usually does some pretty boring backgrounds, I wanna see if I can change that

Hmm, can make some interesting images but no character. Either I should use a different model or the embedding is overfitted...

finally an interesting background.. but the character doesn't look too great

that's actually pretty good

So you still get creative freedom over the character it seems, just not over the background as much.

Looks so good upscaled

[wispy nest]

what does vae stand for?

[wispy nest]

Almost perfect but again hands problem and one eye was a bit out of symmetry so I covered with a pair of glasses

[thorny plaza]

Hey everyone!
I've been playing with SD for a little while now. And I was wondering if it was possible to keep the structure of an image and only change its look. I'm planning to do that on this CAD render :
Is there any way to do that inside stable diffusion, should I use Dreambooth or any other extension ?

[open sun]

Let us all know our corporate slaver "heroes". Good luck giving a guy like this say in art law matters.

[wispy nest]

when surveillance is for your safety

[sharp merlin]

the NSA wants to know your location

[vague oasis]

depth2img could help

[wispy nest]

both the image and the music generated by opensource ais

[desert jolt]

did a few slideshows https://www.youtube.com/watch?v=Nqlmu-lEYxM&list=PL8G7-J1OZZnTBuWrapRm8HMKfR_gHeDja&index=1

[serene inlet]

Some guy DMed this. Is this a scam????

[vague oasis]

i also got that message and joined, its not a scam but its also not the best way to ask someone to join their server

if its free then we are the product, i assume they want to develop a custom model like midjourney does

[serene inlet]

Mods ended up kicking/banning the guy.

Well the server they advertised had like 20K users allegedly. But I was suspicious because I’ve seen some scammers try this before.

[vague oasis]

"Why is BlueWillow spamming messages to users on other AI image generator discord servers?
We hope this isn’t happening! We did setup a reward for getting new users so this might have something to do with this."

from their faq

i dont believe them tho but who knows

[serene inlet]

🤔

[fresh hound]

advertising your own channel in someone elses is shitty imo

especially when its trying to take their "customers" so to speak

Im sure SD wants people to use SD and not someone elses!

[wispy nest]

Nike jordan 4

[broken oracle]

@tropic shell Landon is soliciting people to join his discord server

[tropic shell]

F u n. Thanks!

[tropic shell]

Can you file a ticket with the username and number, please?

[broken oracle]

What channel would I do that in?

[tropic shell]

#1010934719455707218

[hasty nova]

I got one of those too, but since I have to accept message requests I just hit ignore and block, so I don't know what their username was

I watched this, it was a good video

[hasty nova]

A very large crown

Another very large crown

This is nice

Oh this is better

[hasty nova]

Abigail from SDV

these are so fun

[gritty cedar]

looking good

[hasty nova]

Thank you!

It's an embedding i trained a while ago on Abigail from sdv, and I am combining it with my ghost embedding. it kind of gives it a watercolor look for some reason.

[brittle wasp]

@small hearth Some guy is trying to rip off Dreambooth on the Play Store.
https://play.google.com/store/apps/details?id=com.Dreamapps.Dreamboothaiimagegenerator&hl=en_US&gl=US
Developer email: [e-mail]
Kuru got hit with NAI too.
If you want to report it, here's a link.
https://support.google.com/legal/troubleshooter/1114905

[cloud solstice]

Didn't know where to put this, but what's the best way to take an image like this and make the AI create this creature doing different things like holding weapons or something? I want it in a different style I just created it in Blender real quick because the AI has a lot of trouble making this kind of creature.

[small hearth]

It’s fine dreambooth isn’t us

[brittle wasp]

Oh LOL, I forgot the difference

[desert horizon]

@pallid relic i have this

and change the model at top left

@pallid relic i put it here ?

[pallid relic]

@desert horizon the batch file or the embedding?

[desert horizon]

both ? i dunno

[pallid relic]

heads to #🤝｜tech-support

[desert horizon]

like I said , I got hepl with this setup last year in a DM he just sendt me links

[wispy nest]

NSA always knows my location

[wispy nest]

is a new model possible this week?

[coral remnant]

Hey! I tried to make a rigged model with asset done on SB, that's really bad but that's also my first try with live2d. With a bit of work we can get something nice i guess. Still, it prove me that this is a possible way to go from SB to animation without any talent nor knowledge in few hours. Just want to share it

[storm timber]

i have a pub. i need pictures for my instagram account.

[eager heath]

768 vs 1024px

to clarify, the second image is 768px

They clearly look more vibrant in terms of color, but they also feel richer when you look at them more closely

Same number of steps

[fringe sandal]

they both look pretty similar in terms of style, also a cityscape i feel has more photos of it in 768 and 1024 res than a nature landscape does

[eager heath]

Now with more steps, 20 instead of 12. It improves the images a lot but I can't shake the feeling they aren't quite as good as the 12 step 768px ones

Maybe you just need more steps at higher res for the same quality

Because they definitely look more "real" now

But check same steps (20) and lower (768px) res

seed and prompt and other settings never change of course

25 samples, 1024

I still like the 768px results better, but I think more samples improves it

[steel yarrow]

That's what I always say.

why would anyone build a light house in the middle of the mountains anyway

[gloomy elbow]

I thought the text explained that clearly.

[steel yarrow]

... yeah... exac... exactly

[idle spruce]

it's so weird how crisp and clean the text is

but how wrong it's all put together and how it melds some characters together.

it's alien.

i hope we get out of this uncanny valley with AI soon enough.

might take a few years.

might take a new method that completely rethinks how AI learns from images or generates its images.

[steel yarrow]

it even made a new letter, it's like a J and U smashed together

[dry crow]

When deepfloyd is released it will be a lot better i think

[steel yarrow]

how about that pix2pix

turn them to dogs

[golden umbra]

So far from my testing 2.1 feels kinda mid, the laxpeint embedding saved it though since it's the exact aesthetic I'm into

[spiral orbit]

i posted a bunch of reference images on 4chan's SFW request board and asked them over on /g/ in the SD thread if someone could train a model based on those images and the post some outputs. The results have been pretty cool. If you want to give it a shot, the reference images are on this thread: https://boards.4channel.org/wsr/thread/1309891

[wispy nest]

maybe they know something we don't? ;P

[strange dirge]

#1014939219904450590

[acoustic eagle]

@wispy nest what is blurry about this image it made me? https://media.discordapp.net/attachments/1065969190080614440/1067068920231636992/technium_pikachu_as_a_cyborg_android_robot_from_cyberpunk_netru_bdae82a7-d3e5-45b7-9736-013c21f837bc.png

[hasty nova]

Somewhat close to MJ, not exactly though ^^

[acoustic eagle]

I would say that this is a common style of MJ

[wispy nest]

for me, those highlighted white areas are too blurry for my taste. Also, I don't like the "depth focus" in most images. But it's still a great image! Nice! :D

[hasty nova]

It's an image from a randomly generated prompt that a prompt generator me and someone else are working on made

[hasty nova]

Was "porcelain" in the prompt for that?

[acoustic eagle]

This is closer to what sd 2.1 produces me. Incoherent nonsense without major work and many many gens

This prompt was "Dog" with 75 steps

[hasty nova]

I haven't moved on to 2.1 yet

[acoustic eagle]

ah, I think you'd be dissapointed

[hasty nova]

I can try it out though for dog on 75 steps and send the first thing i get

(2.1 uses much different prompting where you have to be more detailed and use the neg to your advantage)

[acoustic eagle]

If I do same prompt seed cfg scale steps with "dog" prompt in 1.5, I get a much more coherent result

still poor

but far better

but let me do "dog" prompt for midjourney, it gens these

[hasty nova]

SD prompts much differently from MJ

I can try to get results similar to those above, I will se what I can do

In 2.1

Which I don't use often

[acoustic eagle]

This is with openjourney model for SD same settings:

still no where near

with protogenxrealism :

lol

umbrellahead will be next scp horror

[hasty nova]

Getting there, will try a few more with a slightly different prompt

[wispy nest]

MJ is better for most people, you can get "better" results using your own local stuff but it needs a lot more work and it's not a certainty it'll be "better" in the end :P

[acoustic eagle]

this is with dog as prompt in 2.1?

[hasty nova]

No, as I said, SD prompts much differently than MJ. For all I know, MJ could add on extra words to prompts to make them have a cooler style. I made a decent prompt for 2.1 (without ever using it really) and that image came out first try.

[acoustic eagle]

can you share prompt and seed?

Is my 2.1 model file corrupted maybe?

[hasty nova]

I'm using a 2.1 model called primevalNova

I am going to try same prompt on normal 2.1

[acoustic eagle]

I cant find the model on google :<

[hasty nova]

lemme grab the link for ya

https://civitai.com/models/4727/primeval-nova-v2

[acoustic eagle]

hmm a very new model

dl'ing

[hasty nova]

👍

I am trying out my prompt on normal sd as well

A little cropped, nothing a little more negative prompting can't fix

[wispy nest]

and this is one of the styles I like. It probably doesn't have as many details, but I still like it :)

[hasty nova]

Oh that's so cute!

[acoustic eagle]

okay lemme get your steps/cfg/seed I want to see if my sd2.1 is corrupted

[acoustic eagle]

what model is that?

sd2.1?

that one is pretty close to MJ imo

[wispy nest]

no, it's a mix of 1.5 models I... well, mixed together :P

[hasty nova]

parameters

Dog, ethereal, fluffy, majestic, cinematic, portrait, close up, hyperdetailed, photogenic, glowing, head, side-profile, fairy dust, light rays
Negative prompt: Noise, background, pop art, cartoon, digital, lines, grain, distortion, ugly, mutated
Steps: 75, Sampler: Euler, CFG scale: 9, Seed: 2230194709, Size: 768x768, Model hash: ff144a4984

[acoustic eagle]

you should share it to civit

[hasty nova]

I apparently have the 5000th upload to civitAI if the number in the link is the number of when it was uploaded

With my 1.5 embed from a few days ago

[wispy nest]

why it has no meta data in the image, and why I don't show it, is because some of the mixed models are NSFW and I don't want to assume it's okay to link those things :)

[acoustic eagle]

You know what maybe was my problem?

[hasty nova]

That's fine here I am pretty sure, people post non nsfw things with nsfw models all the time

[acoustic eagle]

I was using 512x512 for 2.1

It may have ruined the gen

[hasty nova]

Yep...

That's an issue if you are using the 768 model

[acoustic eagle]

mmm

well

[hasty nova]

Now, trying just the word "dog" will still get you messed up results

[acoustic eagle]

Im getting an error trying to load primeval

I'll try going back to sd2.1

[hasty nova]

Did you set up the .yaml?

[acoustic eagle]

oops 😄

2.x too complicated 😭

[acoustic eagle]

I got this silly image from another model it defaulted back to tho 😄

[wispy nest]

And if I'm gonna be honest, then I don't think 2.1 is that "good" compared to the hundreds, if not thousands of side models, etc that 1.5 has so far :/
But then again, I just like it when people can make awesome art! :D

[acoustic eagle]

well look at that

SD2.1 made a coherent image for me now lmao

[hasty nova]

Wooo!

Gald we figured it out :)

[hasty nova]

I got this silly little pup :)

eyes are a little funky

and so is the mouth

but I love it regardless

[dense tapir]

Who trains embeddings in here?

[hasty nova]

I've only done one for 1.5

ethereal pupper, like MJ

[wispy nest]

and I don't know how "good" MJ is at recreating the same character, or a similar character, in different poses, but then again, I haven't even tried to look it up. I've only seen the awesome images people have shared :D

[dense tapir]

I am trying to figure out what to stick on the prompt for my train epoch images.

[hasty nova]

Maybe ask in #🔧｜finetune ?

[dense tapir]

Colab trainer is down at the moment as it is missing a file so TLB has to fix that.

[hasty nova]

I'm not very knowledgable when it comes to this stuff

[dense tapir]

finetune is normally Dreambooth related

[hasty nova]

Then i'm sorry, I can't help you :(

[dense tapir]

No worries. I have done a lot of embeddings and release a bunch too but I never used the prompt before

[hasty nova]

Since you are still on 1.5, would you be willing to try out an embedding I made the other day? It's kind of a spooky ghost one, but using it just as a style can kind of make your images very ethereal. Also really good at portraits. I just really like seeing what other people make with things I have made..

[dense tapir]

I am no fan of MJ

[hasty nova]

I don't like their whoel discord thing

And how it costs money, because I have none

[dense tapir]

That is why. I was introduced to this and never did understand how to use it so I gave up and SD came ou. Here i have been since.

Costing money is BS

give us the tools to run locally or on colab and don't charge.

[wispy nest]

sure, I'm a little busy at the moment, but if you want, then I can always PM you with with the results etc later, or you can always PM me with the embedding etc :)

[dense tapir]

MJ makes most of its money from their clients anyway

[hasty nova]

Ah, it's all good, if you are busy I won't bother you with it