初嘗 Stable Diffusion

起初以為用 Stable Diffusion 可以直接打一句 prompt 就可以出到想要嘅嘢,就有如係 ChatGPT 咁問一條問題或者俾一個指示,玩左幾下發現原來要有好少少嘅圖,係要俾一個幾清晰嘅指示。好似畫一個空倉庫咁,係參考完唔同嘅網上資源之後,寫左下面正面同負面嘅 prmopt ,先至叫做可以出到令人滿意嘅圖。

Positive prompt

High-fashion photography in an abandoned industrial warehouse, with dramatic lighting and edgy outfits,extremely detailed CG unity 8k wallpaper, detailed and intricate, original,highres,best quality

Negative prompt

paintings, sketches, (worst quality:2), (low quality:2), (normal quality:2), lowres, ((monochrome)), ((grayscale)), skin spots, acnes, skin blemishes, age spot, glans,extra fingers,fewer fingers,strange fingers,bad hand,signature, watermark, username, blurry, bad feet,bad leg, duplicate, extra limb, ugly, disgusting, poorly drawn hands, missing limb, floating limbs, disconnected limbs, malformed hands, blurry,mutated hands and fingers

Read more →

安裝 Stable Diffusion

近年人工智能技術急速發展,而自從 2022 年出現了 DALL·E 2ChatGPT 3.5 之後, Generative AI 就成為左人工智能界嘅熱門話題。

可惜嘅係,香港自從由亂入治之後,好多新嘅高科技嘢香港都冇得用,而 OpenAI 嘅 service 都唔例外,所以 Dall-E 呢個由交字生成圖片嘅 AI 香港就未必用到啦。不過唔緊要,市場上仲有 Stable Diffusion 呢個 opensource framework 可以用,而我就決定下用 Stable Diffusion 去生產圖片。

Read more →