NokiMo
狂妄画集
狂妄画集

fanbox


胶衣真好看

剩下的问题就是解决表情了和画风了~

胶衣真好看 胶衣真好看 胶衣真好看 胶衣真好看

Comments

In other words, NovelAI doesn't have the ability to have you upload a set of works or images as reference materials for the A.I. to train the general art to your particular style using it's general pool of vast datasets/learning as a basis for reference? Like giving an artist a set of parameters & reference photos to use to show them what you are thinking about or looking for in a final image. 換句話說,NoveAI 沒有能力讓你上傳一組作品或圖像作為 A.I. 的參考資料。 使用其龐大的數據集/學習的通用池作為參考基礎,將通用藝術訓練為您的特定風格? 就像給藝術家一組參數和參考照片,用來向他們展示您在最終圖像中的想法或尋找的東西。

AI对于姿势的学习并不是太好,一般的动作都是要通过绘画和关键字来调试,我一般就是先画好草图基础草图,然后AI自行识别绘制,然后反复调整。 不过一般常规一点的姿势AI还是能够生成的 一些绘画功底比较好的画师,一般会在最后一步改成自己的风格,不过我就有点苦手了

一般来说自己搭建需要大量的作品作为参考,目前还是用的novelai,是已经学习好了的

這些服裝看起來非常性感! 結果他們做得很好! 接下來,也許是人工智能。 可以調整以繪製類似的材料,但樣式更加多樣化,例如戰鬥服或其他變態服裝? 您是否能夠讓該應用程序產生更多動態姿勢? 不同的武術姿勢? 其他動態姿勢? 如果有一種方法可以將自定義的姿勢緩存提供給 A.I. 在繪畫之前,也許可以用動態的打鬥姿勢和麵部表情,或者性感的姿勢和麵部表情來調整圖像? 如果沒有,那麼有沒有辦法剪切和粘貼可以滿足這些標準的特定線條藝術? 我已經看到其他藝術家做了類似的事情,並成功地將他們自己的藝術複製到了令人不安的準確程度! 😅

The outfits look EXTREMELY sexy! They turned out quite well! Next, perhaps, the A.I. can be tuned to draw similar materials, but in a greater diversity of styles, such as battlesuits or other kinky attire? Have you been able to get the app to produce more dynamic poses? Different martial arts stances? Other dynamic poses? If there is a way to feed a custom cache of poses into the A.I. prior to drawing, then perhaps images of dynamic fighting poses and facial expressions, or sexy poses and facial expressions can be used to tune the image? If not, then is there a way to cut & paste particular line art that could meet these criteria? I've seen other artists do something like this to great success replicating their own art to a disturbingly accurate degree! 😅

⬆️ Machine translation for ease of reply for non-english speakers. :3

有什麼方法可以養活人工智能嗎? 一組自定義圖像來訓練它以進行特定的圖像生成嘗試? 也許,粘貼您正在尋找的已知面部表情的特定圖像來幫助“訓練”人工智能? 或者,也許,像我們練習畫面部表情一樣嘗試畫簡單的線條表達? 🤔 我知道一個練習,你在一張紙上快速連續地為眼睛畫兩個點,然後在嘴巴上畫一條線。 通過超快地完成它,您可以獲得大量的變化,這些變化可以幫助您隨著時間的推移創建面部表情目錄。 另一個想法可能是擁有一本書或一系列面部表情線條藝術,然後僅將表情複製到預生成的圖像上,以幫助調整人工智能。 我已經看到其他藝術家在繪製人類陰莖時取得了巨大的成功。 😅 有助於消除畸形或其他異常。

Is there any way to feed the A.I. a custom set of images to train it for a particular image generation attempt? Maybe, paste a particular image of a known facial expression you are looking for to help "train" the A.I.? Or, perhaps, try to draw simple line expressions like we do to practice drawing facial expressions? 🤔 I know of one exercise where you draw two dots for the eyes and then a line for the mouth over and over in quick succession on a sheet of paper. By doing it super fast you get NUMEROUS variations that can help you over time create a facial expression catalog. Another idea might be having a book or collection of line art of facial expressions and then copy just the expression onto the image pre generation to help tune the A.I. I've seen other artists do this to great success when drawing the human phallus. 😅 Helps eliminate deformity or other abnormalities.


Related Creators