1024,のTwitterイラスト検索結果。 33 件

ssd-1b-anime-v2
ssd-1b・・・扱いが難しくない！？！？
SimplepositiveXLv1 1girl
Negative prompt: negativeXL_D sdxl-negprompt8-v1m unaestheticXL_AYv1
Steps: 20, Sampler: Euler a, CFG scale: 7, Seed: 1, Size: 1024x1024, Model hash: 4c5456f566, Model: ssd-1b-anime-v2
12.8 sec

同じ条件でsd_xl_base_1.0_0.9vae
う、うん・・・！だよね！！
SimplepositiveXLv1 1girl
Negative prompt: negativeXL_D sdxl-negprompt8-v1m unaestheticXL_AYv1
Steps: 20, Sampler: Euler a, CFG scale: 7, Seed: 1, Size: 1024x1024, Model hash: e6bb9ea85b, Model: sd_xl_base_1.0_0.9vae

SSD-1B（SDXLの軽量版実験）
なんかえらいこっちゃになったな・・・
SimplepositiveXLv1 1girl
Negative prompt: negativeXL_D sdxl-negprompt8-v1m unaestheticXL_AYv1
Steps: 20, Sampler: Euler a, CFG scale: 7, Seed: 1, Size: 1024x1024, Model hash: 0bf1ce6b06, Model: SSD-1B,

WebUIの動作報告
i7-10875H
RTX 3060 Laptop (VRAM 6G)
ARGS=--xformers --disable-nan-check --medvram --always-batch-cond-uncond --opt-channelslast --opt-split-attention-v1
1024×1024,bat size = 8で8.3s/it、9で9.5/itぐらいの速度感

biya1024, for fapping purposes

BPModelついて。
少し触ってみましたが、「優秀な問題児」という感じでした🤔
ACertaintyを高解像度で学習したということで、高解像度でも破綻が少ない気がします。
一枚目からhighres. fixなしで960x640, 1152x768, 1536x1024, 1920x1280です。
https://t.co/kXWPXmxx5e
(1/4)

Steps:30,Euler-a,768x1024,Batch-count:10,Batch-size:1,Scale:7.5,Anything
で10枚で181秒でした
これを基準にRTX4090に差し替えて何秒になるか、楽しみですねえ
あと、メモリ足りなくて解像度やBatch-size増やせないとかもあるので、それが改善されるとなると、うっ！！！

ボツ絵の供養。 #StableDiffusion

MODEL: v1-5-pruned-emaonly.ckpt
Steps: 100, Sampler: Euler, CFG scale: 11
Seed: 24805162, Face restoration: GFPGAN
Size: 1408x1024, Denoising strength: 0.6, First pass size: 704x512

Demonic Ecstasy, 2022
Stable Diffusion, Topaz Video AI
1024x1024, 72.7MB, MP4

I aim to capture beauty and decay, gentleness and horror. There is an agony that I feel was captured here, but also a release, a rebirth. Not an end, but a new beginning.

#darkart #horrorart

.@luadsix 『2,4,8,16,32,64,128,256,512,1024,2048,4096,8192……』

1024,576,24,1337,0.557 #AI #NovelAI #NovelAIDiffusion #PureNegativePrompt https://t.co/yTDXU2Y91I

W512,H768 or 1024,Wuler aの生成例
modelはwaifu-diffusionとtrinart2をmerge
token数はWebUIで見れたやつを書いてますが、25~30に抑えてます。（というよりtxt2の仕組みに詳しくなくてこれ以上付加の形でいじれない）
3060laptopだと、裏でブラウザが動いていても3分20秒程度でこれが出てきます。

optimized stable diffusionやってみた。
順に一辺256, 512, 1024, 1536pixel。
#stablediffusion #aiart

Pixel Style Animated Gif (640*1024, 800%)
ちびキャラのドット絵風アニメーションを描いてみました
#葵色画廊 #ほめろゆげまろ
(1/3)

'THE TEMPLE'

The stitching part of @huemin_art's JAX 2.7 is sort of groundbreaking to me, all the details! Started at 512x1024, it's now at 6114x12228. Click the link to see the full version

https://t.co/gIOpCT8mFm

#aiart #artcommunity #digitalart

Prompt: ocean sailing, highly detailed matte painting, storm, lightning, by Josan Gonzalez:2

use a gold > teal > turquoise color Gradients as init_image

size 1024, CGS:6000, Step:1000, Skip:step:300

post editing: Cthulhu's hug added 😂

#aiart #discodiffusion

"AZ Tower, ATZ1024, holding short 17R, ready for departure." 🛫
#ATEEZFanart #MINGI

Here's the model with the full-res, filtered textures applied.
The map was originally 1024x1024, but I sized it down to 256 and indexed the palette to 32 colors for extra cronch. Makes her look a bit more like a character in an early-00's PC game
#lowpoly #indiedev #gamedev

Stack (6x9) minting on @fx_hash_ in 2h20m. My first collection there! Edition of 1024, 5tez #xtz #tezos #cleannft #generativeart #fxhash

En el #WallpaperPSS de hoy:

❤️‍🔥 'Cold Game' de Kaneyoshi Izumi
🗓️ Revista Betsucomi (Shogakukan), n.º 02/2022

📱 1080x1920
💻 1280x1024, 1920x1080 y 2160x1080