Can a free AI image model spell? (SDXL text test)

0 Просмотры

  • Спасибо! Поделитесь с друзьями!

    Вам не понравилось видео. Спасибо за то что поделились своим мнением!

aii
Издатель
I asked Stable Diffusion XL for 3 easy words, 2 seeds each:
"OPEN" on a shop sign: ✅ the neon sign is right ❌ the wooden sign above is unreadable · seed 2: "OPIEI"
"HELLO" on a coffee cup: ⚠️ "hello" (almost) · seed 2: "Helo"
"HAPPY BIRTHDAY" on a cake: ❌ "BAPPY", "PIRTID" · seed 2: candles "1 8 1", no words
Score: 1 word right out of 6 images.

Need text? Add it yourself afterwards in any free editor: 10 seconds, always right.
Settings: 1024 px, 25 steps, CFG 5, negative prompt "blurry, deformed" (no "text", so it was allowed to write).

Tools: Stable Diffusion XL base 1.0 by Stability AI (CreativeML Open RAIL++-M: commercial use allowed, with use-based restrictions), run locally with Hugging Face diffusers (Apache-2.0).
Narration voice: PocketTTS by Kyutai (CC-BY 4.0). Captions: Whisper large-v3-turbo via faster-whisper (MIT).

#AI #StableDiffusion #SDXL #AIArt #AIFails
Категория
AI Картинки
Выполните вход или зарегистрируйтесь для отправки комментария.
Комментариев нет.