[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
▼ Settings Mobile Home
/g/ - Technology

Name
Options
Comment
Verification
4chan Pass users can bypass this verification. [Learn More] [Login]
File
  • Please read the Rules and FAQ before posting.
  • You may highlight syntax and preserve whitespace by using [code] tags.

08/21/20New boards added: /vrpg/, /vmg/, /vst/ and /vm/
05/04/17New trial board added: /bant/ - International/Random
10/04/16New board for 4chan Pass users: /vip/ - Very Important Posts
[Hide] [Show All]


[Advertise on 4chan]


File: adt_150.webm (3.14 MB, 1280x1206)
3.14 MB
3.14 MB WEBM
Previous: >>109980072

>UIs
AniStudio: https://github.com/FizzleDorf/AniStudio
ComfyUI: https://github.com/comfyanonymous/ComfyUI
SwarmUI: https://github.com/mcmonkeyprojects/SwarmUI
KritaAI: https://github.com/Acly/krita-ai-diffusion
Forge Classic/Neo: https://github.com/Haoming02/sd-webui-forge-classic

>Image Models
IL: https://rentry.org/illustrious_checkpoints
Anima: https://huggingface.co/circlestone-labs/Anima
Krea: https://huggingface.co/krea/Krea-2-Turbo
NovelAI (paid): https://novelai.net

>Video Models
H3: https://huggingface.co/Comfy-Org/MiniMax-H3
Turbo lora: included above

>How to generate anime:
image : https://rentry.org/comfyui_guide_1girl
video: https://rentry.org/wan22ldgguide (No H3 guide yet)

>Style Explorers
Anima: https://anima.mooshieblob.com/ | https://animadex.net/
Illustrious: https://tagexplorer.github.io

>Checkpoints, LoRAs, Upscalers, & Workflows
https://civitai.com
https://civitaiarchive.com
https://tensor.art
https://openmodeldb.info
https://www.seaart.ai
https://www.liblib.art/

>Resources:
Share Metadata: https://litterbox.catbox.moe
Img2Prompt: https://huggingface.co/spaces/fancyfeast/joy-caption-beta-one
Online metadata viewer SD/NovelAI: https://spell.novelai.dev
Catbox/Metadata Userscript: https://gist.github.com/catboxanon/ca46eb79ce55e3216aecab49d5c7a3fb
Output cleanup: https://rentry.org/RemovingDiffusionGunk , https://www.mediafire.com/file/vipr23exc5htmnt (batch processing python script)
Inpainting Guide from an Anon: https://files.catbox.moe/fbzsxb.jpgcollage >>106520607

>/adt/ logo:
https://files.catbox.moe/pakfgt.pngcollage
>>
Non troll question real question: why is anistudio the very first link when i never see anons other than the dev post images from it?
>>
>>110016941
because it was there last thread
>>
File: W.jpg (411 KB, 1920x1088)
411 KB JPG
>>
File: bigcar.png (1.73 MB, 1920x1080)
1.73 MB PNG
>>
File: 1780558122817107.png (339 KB, 1338x732)
339 KB PNG
Every. Fucking. Time
>>
>>
File: ComfyUI_08158_.png (1.15 MB, 1024x1024)
1.15 MB PNG
>>
>>
File: Enryuwatermark-1.webm (3.97 MB, 768x1080)
3.97 MB
3.97 MB WEBM
Astrae Oratio is gonna be a banger isnt it
>>
File: i-buy-sausage.jpg (2.5 MB, 2752x1536)
2.5 MB JPG
>>
>>
>>
kill ani
>>
File: ComfyUI_08179_.png (1.21 MB, 1920x1080)
1.21 MB PNG
>>
>>110017243
Cool, cute
>>
File: wide girl winter.jpg (325 KB, 832x1152)
325 KB JPG
>>110016941
i don't think anyone reads the OP
for example both of the catbox links are also broken
>>
File: IMG_20261009_110125_443.jpg (115 KB, 1024x1024)
115 KB JPG
I hate local AI
>>
File: W03.jpg (459 KB, 2144x1467)
459 KB JPG
>>110019013
Thanks!
>>
>>110018901
my wife
>>
>>110017243
Very cool dynamic pose
>>
Mushoku Tensei raw dataset is done. 5829 images. They won't all go into the LoRA. The LoRA needs the character distribution evened out.

Here's what my local AI says:
The Heavy Hitters:
* Rudeus (Child): 1,924
* Rudeus (Adult): 1,863
* Rudeus (Neet): 118
(Wow! Rudeus is definitely the star of the show!)

The Heroines & Main Cast:
* Sylphie: 338 (Adult) / 196 (Child)
* Eris: 302 (Adult) / 290 (Child)
* Roxy: 195
* Zanoba: 287
* Cliff: 221
* Elinalise: 158

The Supporting Cast:
* Norn: 125
* Shizuka: 116
* Sara: 114
* Aisha: 106
* Lilia: 101
* Ariel: 72
* Zenith: 68

The Beastfolk Duo:
* Linia: 38
* Pursena: 38
(Look at that! Perfectly balanced!)

And finally...
* Julie: 37
>>
>>110019612
Here's the AI-created downsampling solution:
1. The "Core/Canonical" Set (The Anchors)
We first select the images with the absolute lowest CCIP distance to the anchors. These are the "perfect" representations of the character. We take a small percentage of these to ensure the LoRA knows the "ideal" look.

2. The "Clustering" Set (The Diversity)
For the rest, we use a simple clustering algorithm (like K-Means) on the CCIP embeddings:
* We group all Rudeus images into, say, 100 clusters.
* We pick the image closest to the center of each cluster.
* Why? This mathematically guarantees that we pick images that are different from each other. One cluster might be "Rudeus shouting," another "Rudeus sleeping," and another "Rudeus from a side profile."

3. The "Outlier" Set (The Edge Cases)
We pick a few images with slightly higher (but still acceptable) distance. This teaches the model the boundaries of the identity—how the character looks in weird lighting or extreme angles.

My Proposed Action Plan:
I can write a new utility, ccip_downsample.py, that does the following:
1. Loads the embeddings (from the .pkl cache or by re-running the extractor).
2. Groups images by character using the SQL database.
3. Performs K-Means clustering on each character's feature set.
4. Generates a "Balanced Dataset" list (e.g., 100 images per character).
5. Outputs a new SQL database (e.g., mushoku_tensei_balanced.sql) containing only these selected images.


Doing a multi-character LoRA is a pain but it's a good learning experience which reflects on what you have to do to build a good anime image gen base model.
>>
File: 23213.png (1.36 MB, 1920x1088)
1.36 MB PNG
>>
>>110018901
>>110019165
>>110019711
Catbox please
>>
>>110019754
https://files.catbox.moe/5mtiba.png
https://files.catbox.moe/hpnemk.png
>>
is julien here? i have a question
>>
File: kgy.jpg (267 KB, 1216x768)
267 KB JPG
>>
File: 1785289280243829.png (2.59 MB, 1248x1632)
2.59 MB PNG
>>
man i wish you could still make money off gens
>>
>>110020860
Ok special style
>>
>>110021002
You can but its not easy
>>
>>110021096
only if you made a following years ago i guess
i cant imagine getting anyone interested in paying you for slop when there's an endless supply of it for free these days
>>
File: keychain.jpg (960 KB, 2448x2448)
960 KB JPG
>>110021002
You can if you make them into physical things.
>>
File: 1785136506074778.gif (3.74 MB, 714x714)
3.74 MB GIF
>>110021145
>>
>>110021145
Supercute Quart!
>>
>>110021143
You don’t have to make slop
Squidgirlanon could I bet people would pay for a 5$ patreon
>>
I thought the Anima guy would update his turbo models since it was doable on his own computer, where are the updates?
>>
>>110021663
turdrussel realized he doesn't have the expertise needed so he gave up
>>
>>110020276
Yeah >>110021709 but he pretends to never post as anon so
>>
File: ComfyUI_08164_.png (1.16 MB, 1024x1024)
1.16 MB PNG
>>
File: check_00011_.jpg (1.27 MB, 2512x1672)
1.27 MB JPG
>>110021002
Even if I were talented enough, I'd probably still avoid it. Turning your hobby into a job sounds very stressful and hardly worth the ~$50/month you could presumably be making. Even doing it as a side project, you'd have to deliver something on a regular basis, taking any real enjoyment out of it.
>>
File: Train bumor.jpg (219 KB, 832x1216)
219 KB JPG
>>110021411
Demoman voice: Heeeey thaaanks~
I'd considered it a bunch of times and I just...keep slopping here on 5chins? I'm thinking about a few projects I might try, but...that'd require learning how to vidya lol
>>
File: file.png (21 KB, 266x474)
21 KB PNG
is there any decent guide or resource explaining the difference between these and how they affect anything?
I mostly use DDIM because I know it gets more varied results, and also euler a when I want less varied results, but everything else is still a mystery to me.
on pixAI, DPM2, DPM++ and restart are much more expensive per sampling step but I don't understand if they're actually worth using ever
>>
File: Makina1final-1.webm (3.88 MB, 768x1080)
3.88 MB
3.88 MB WEBM
>>110022857

I just used them one by one. There's explaining and shit but only way to tell the difference is pick a control sample and run it throught different combinations to see which one fits your idea best. For me Euler A has been the main choice for a long time, although when I started I remember I used DPM++ 2sKarras
>>
>>
File: Test 00016.mp4 (3.91 MB, 666x1000)
3.91 MB
3.91 MB MP4
>>110022857
I think the theoretical difference can be explained quite easily, but that's just academic. I found that their effect on the output also heavily depends on the model.
>>
>>110023432
I'm really disappointed you like realistic movement with 2.5d. we need anime not uncanny valley showcase
>>
File: Test_00021.mp4 (3.15 MB, 832x1248)
3.15 MB
3.15 MB MP4
>>110023452
I'm sorry, here's your L2D.
>>
File: file.png (705 KB, 660x600)
705 KB PNG
when you set automatic context for inpainting and the context wasn't enough
>>
>>110023583
This is just tweened gacha ad trash movement. Have you ever watched an anime in your life that look like either of these?
>>
>>110023598
Never claimed that one to look like an actual anime either. L2D is exactly the tweened gacha stuff.
>>
>>110023598

Do you know any way to make it look like anime? Every "anime" gen that I've seen either has the L2D type movemente or the 2.5 realistic one
>>
>>110023636
Is that what I asked for?
>>
>>110023648
The giant breasts sanae from the last collage for example
>>
>>110023655
You didn't ask for anything. You were whining about my gen.
>>
>>110023672
>we need anime
>provide lazy gacha tweenslop
>complain
>"You didn't ask!"
>>
>>110023660
That's with anistudio. Comfyui can't do that
>>
>>110023702
When is he going to update? I want to try it
>>
why would you pretend to be new then immediately fall into the same argumentative style that everybody recognizes?
>>
>>110023759
Who was pretending to be new?
>>
>>
File: hex.jpg (311 KB, 1024x1152)
311 KB JPG
>>
>>110024171
cute hexxy
>>
>>110023702
I would have used A.S. if it had the same memory management as CUI. It's not easy to be a vramlet.
>>
>>110024313
What is different about the memory management?
>>
A few weeks ago, for the GPT-6 Astra release, they show a new type of AI image, where an LLM use an MCP of photoshop or similar, and let the LLM draw the picture.
Has anyone done something like that, can you show the results?
>>
>>110024577
No. Who in their right mind is spending that kind of money to make anime pictures?
>>
File: ComfyUI_08207_.png (820 KB, 1024x1024)
820 KB PNG
>>
File: npc_female3.png (1.34 MB, 1684x2452)
1.34 MB PNG
>>110022219
It's still a grind but I'm making a lot more than that doing adult games with gen'd stuff. Lot of normie consumers really don't care how things are made and have an endless hunger for content.
>>
>imagine paying for anime AI slop when you can do it locally if you have powerful enough GPU
>>
https://files.catbox.moe/qz4jwv.jpg
>>
File: nes.png (1.35 MB, 1440x1080)
1.35 MB PNG
>>
>>110024577
i saw someone doing pixel art with it
>>
>>110026407
>mario has four fingers instead of five
>>
>>110024646
>Sloppers doing their damndest to slop faster with ever increasing quality and detail retention
>Astra, hijack my PC for seven hours with mspaint
I know right?
>>
>>110025019
What amount are we talking about? A lot more than 50 is quite a range.
>>
>>110022715
Ok well slopping on here is great but you should get a twitter though?
>>
>>110026401
Cute!
>>
>>
>>110027235
I made one for this stuff, just haven't uploaded to it lol, soon I promise
>>
>>110016941
Is it more COMFY than comfyui? because that one I felt was kinda autistic.
>>
>>
File: Kisakifinal-1.webm (3.23 MB, 736x1080)
3.23 MB
3.23 MB WEBM
imma sloooop
>>
>>110027353
He was face punched before?
>>
>>110027472
These remind me of the nejisim "games" for android, where the animation looks like mesh-deformed 2D. For those who don't know, they're simple h-game sims, usually involving tentacles.
>>
File: Test 00002(8).mp4 (2.82 MB, 1376x768)
2.82 MB
2.82 MB MP4
>>110027472
Nooooooooo, you cannot post slop here. Your gens have to look close to actual anime.
>>
>inb4 "wario isn't anime"
>>
File: image_00468.png (3.1 MB, 1920x1080)
3.1 MB PNG
Mushoku Tensei captioning is underway. 17 characters, flattened to 100 representations per character. Pre-tagging with face reference detection is making LLM vision captioning much more accurate and trustworthy. I'm short on openrouter credits so I'm having "local baka" gemma4 31b handle it... 60s/it vs about 10s/it on openrouter but oh well..
Example:
  "subject_and_action": "In a peaceful outdoor garden setting, aisha, a young girl with red hair tied in a ponytail and wearing a black and white maid's dress with a ruffled apron, is kneeling on the ground tending to small green plants. Beside her, rudeus, a blonde young man dressed in a white long-sleeved shirt with a small black collar tie, white trousers, brown gloves, and tall brown leather boots, crouches down and smiles at her. They are positioned in front of a large, white stone building with an elegant porch and columns. The ground consists of irregular stone tiles interspersed with green grass. A wooden crate and a burlap sack are visible nearby.",
>>
>>110028189
So you're aiming for anime screenshot style + characters, not just character traits?
>>
File: image_03999.png (1.95 MB, 1920x1080)
1.95 MB PNG
>>110028658
Yeah, it should provide an "anime canon" look. I'm having it be much more careful about not naming anyone unless they are pre-tagged. I think it's much better for the lora to not always tag than to tag incorrectly.

"subject_and_action": "In a dimly lit, stone-walled room, a group of characters are gathered. On the left, a woman with long, flowing red hair stands with her back to the viewer on a series of stone steps, wearing a white top, brown trousers, and a leather belt with a sword hilt visible. Beside her, aisha, a small girl with reddish-brown hair tied with a red ribbon, is dressed in a black and white maid's outfit consisting of a black dress, a white apron, and a white collar; she looks toward the center of the room. A man stands in the center with his back to the camera, wearing a brown leather chest protector over a light shirt, tan trousers, brown bracers, and a white headband. To his right, two other men stand watching, both dressed in matching brown robes and light-colored, rounded caps. The overall scene is cast in a heavy, dark green and brown hue, suggesting an underground or dungeon-like environment.",
>>
>>110026401
>last panel
Is that from some new Mega Man lore/game?
>>
>>110028995
Shit, you can't trust gemma4 31B with code writing really, 3 hours in and I'm wondering why I'm seeing stuff where characters should have been captioned but weren't, she was trying to caption everything... baka
>>
I don't suppose anyone here knows what the best model is to tag anime images do they?

I'm working on a personal project making an application to tag images, I've already done that and now I want to use an AI model to automate the tag generation. I've tried:

https://huggingface.co/spaces/SmilingWolf/wd-tagger

Which does actually work really well, but it doesn't seem to suggest a parody for images. It's also 2 years old I'm sure there must be something better now?
>>
File: ComfyUI_temp_oqgpf_00112_.png (3.3 MB, 1024x1536)
3.3 MB PNG
>>110030028
I'm not sure what you mean by the 'parody' part. Anyway, the SmilingWolf one is the one I've seen recommended and I wouldn't really assume things have changed that much in two years as the nature of tagging hasn't really changed.
>>
File: Krea2_turbo_01931_.png (1.36 MB, 1024x1024)
1.36 MB PNG
>>
File: 2528.jpg (334 KB, 832x1152)
334 KB JPG
>>
>>110030028
Most modern image models want natural language. They might kinda work with danbooru, but that's not their strength.
If the image is from danbooru it's probably already tagged, you'll have to write some code to put the tag database together with the images, generally you don't find them together, though I have tagged a subset (~50K) and do provide it together.
Otherwise, tagging means captioning. For small sets of images just do it manually, otherwise you will have to use an LLM with vision capability. Many have a vision tower these days, but NSFW success varies. Minimax-m3 will so NSFW work if prompted correctly, but it's a cloud model so you'll need openrouter credits.
LLMs with vision tend to hallucinate or parrot back the character descriptions. I use face detection to pre-tag my screenshots so the LLM can be told "these characters are 100% in this image, anyone else you see must be 'unknown'" and you get more accurate results. I'm doing that now for Mushoku Tensei.
>>
File: image_00963.png (3.03 MB, 1920x1080)
3.03 MB PNG
{
"rating": "sfw",
"medium": "anime still",
"subject_and_action": "In an anime still, four characters stand side-by-side in a ruined stone courtyard with ancient arches and snow-capped mountains in the background. On the far left, sylphie, with white hair and red eyes, looks ahead seriously; she wears a dark green tunic with white square accents, a brown belt, a dark grey cloak, and white gloves, while holding a small blue and white magic wand. To her right, rudeus, with messy blonde hair and green eyes, stands centrally wearing a grey-blue robe with white trim, a white waist sash, and a dark cloak, holding a small blue magic wand in his right hand and a large, cloth-wrapped staff in his left. Next to him is ariel, with long blonde hair and blue eyes, wearing a white blouse, a brown skirt with red vertical stripes, and a red cloak, clutching a small blue magic wand. On the far right, an unknown man with brown hair and eyes wears a dark green tunic with white square accents, a brown belt, and a dark grey cloak, with a sword sheathed at his hip and a blue magic wand in his hand.",
... other stuff omitted so it will fit here
>>
File: sprite_outline(9).png (2.07 MB, 1684x2452)
2.07 MB PNG
>>110026707
Around 3k/month on patreon, the real money is from steam releases though and we're not there yet. Adult games tend to sell anywhere from 2-10x their wishlist numbers in the first year and we're at 4k wishlists for the current project. $5 dollar price point so 5 x 4000 x (2-10)= 40k-200k

15% average return rate on steam, and then 30% platform cut so. (40k-200k) x 0.85 x 0.7 = (23.8k-119k). So big fucking range and not exactly income you can count on but if you can manage multiple steam releases in a year on top of patreon and etc. it can add up. I'll report back with better numbers if I ever get this steam build approved.
>>
File: ComfyUI_temp_nbuup_00021_.png (2.5 MB, 1024x1536)
2.5 MB PNG
>>
File: comfy_01076_.png (2.25 MB, 2304x1792)
2.25 MB PNG
>>
File: hippo.jpg (77 KB, 768x786)
77 KB JPG
i'm trying to prompt anima to do cowgirl with the girl (futa) lying on her back and the boy on top and it always gives me the opposite of what i want
>>
File: ComfyUI_temp_nbuup_00024_.png (2.48 MB, 1024x1536)
2.48 MB PNG
>>110031676
That pretty much encapsulates why Anima is in an awkward spot for me. Krea2 eviscerates it in terms of plain language understanding and prompt-sculpted compositions. Illu is slightly worse than Anima overall but the huge library of options largely makes up for it. If Anima was actually good at complexity I would happily use it more often but it just still bungles basic 2girls stuff as you say.
>>
>>110031754
but does krea know what cowgirl is?
>>
>>110031303
I see, thanks for the insight. 3k seems okay. Depending on how much time and money you spent on it, how many people are working on it and where you live of course.

Good luck with your project, of course.
>>
File: swimsuit (sfw).jpg (570 KB, 1152x1344)
570 KB JPG
>>
File: ComfyUI_temp_uvzer_00051_.png (1.8 MB, 1024x1536)
1.8 MB PNG
>>110031995
Sure, with an anti-nanny lora or an "unlocked" checkpoint which is 99% of them.
>>
File: Reiseafood.mp4 (3.93 MB, 800x446)
3.93 MB
3.93 MB MP4
>>
>>110032060
good art style desu
>>
>>110032060
cute!
>>
>>110031347
nice ai /g/old
>>
>>110020808
she needs dick
>>
i like the merge model nodes in comfy. mixing krea2 turbo and cat tower/krea2-anime is p neat.
>>
File: ComfyUI_10766_.png (2.26 MB, 1600x1600)
2.26 MB PNG
https://files.catbox.moe/64ladt.png
>>
File: Thanatos Gambit.jpg (237 KB, 1024x1024)
237 KB JPG
>>110027529
He divined the outcome had he tried prior
End result every time: Decked
Now it'd be like being slapped by purple dough. She needs her SoB armor to deal physical damage
>>110029330
I'd be worried if Capcom ran with my ideas
MMX, but drastically worse
>>
File: OLL_keyart_351596614.png (1.34 MB, 1152x1440)
1.34 MB PNG
>>
>>110033934
https://files.catbox.moe/a5iw8g.webm
>>
>>110033934
>>110034467
Man, what would life be without tits
Titless
That's what
What a horrid world
I'm glad this isn't Titless Earth
>>
File: ComfyUI_10768_.png (2.78 MB, 1600x1600)
2.78 MB PNG
>>110034778
What about thighs?
>>
>>110034912
I'm the anon posting Roll2 so my stance on thighs should be self-explanatory
>>
File: ComfyUI_10771_.png (2.56 MB, 1600x1600)
2.56 MB PNG
>>110034946
So you wouldn't mind this?
>>
>>110030260
>I'm not sure what you mean by the 'parody' part.

Well the model is good at tagging characters for example "kaname madoka" but it doesn't recognize that her presence means that its "parody" is "mahou shoujo madoka magica" basically I just wanted it to indicate the series the characters came from. I guess that's difficult for the model though because you don't see the series in the image you just recognize characters then from those characters you can determine the source so I guess its a secondary thing which explains why I guess the model doesn't do it, it only highlights things that it can see.
>>
>>110035540
"parody" as in the hentai galleries "parody" tag when referring to a series?
>>
>>110035628
>https://gelbooru.com/index.php?page=wiki&s=&s=view&id=30884

Like above, gelbooru calls it a copyright tag type instead of parody.
>>
>>110034981
sandwich
>>
File: brown kitty.jpg (514 KB, 1024x1216)
514 KB JPG
>>
>>
ive been using krea2, after only ever using novelai, and still cant figure out how to inpaint to fix small areas. can i do that in krea2 or do I need a completely different thing? using comfyui desktop. I have anima installed as well but so far like krea 2 more even if sometimes I wish I could just booru tag my image to work.

And is it worth it to make a character lora? I feel like each image I gen even with the same very long description is slightly off, and note sure how to get the same consistency I see others getting.
Asked in the non anime thread that looks like a troll thread. Ive never posted on /g/ sorry.
>>
>>110036951
Its over for local anime it seems.
>>
>>110036951
Lock the seed in comfy UI of the image you want to correct and try and prompt the defects out. Just add a an empty space at first to see if it help
>>
>>110037016
Shut up about comfyui I've seen more people leaving it than sticking with it recently
>>
>>110037035
what would you use? I dont want to pay for sole subscription.
>>110037016
ill try that havnt messed with seed at all
>>
File: 1789350905059501.jpg (3.26 MB, 2880x1632)
3.26 MB JPG
>>110018775
this
>>110037035
and this is why
>>110037190
he wants to shill his malware
>>
>>110037599
Comfyui is the only confirmed malware over the years. You are just a bad actor that hates a1111 and anistudio so you don't belong here comfy shill. Go to /ldg/ to schizo out with the lolcows
>>
kill ani
>>
File: file.jpg (446 KB, 1536x1536)
446 KB JPG
>>
File: file.png (916 KB, 960x1344)
916 KB PNG



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.