The world of AI, Aug, 2024
AI 流行,大量不同程式(不能盡錄),和網上平台可做到,文生圖/圖生圖,出片,語音,文件分析,出報表,作曲/詞,虚擬數字人等,大量的工具和不同應用。而且不停在變化中,需要保持不停學習的心態。
單是生圖,已有不少工具!
自己 PC 安裝 stable diffusion(SD)不是問題,但要快,便要一張不平宜的多 vram 獨顯,幾千元或以上 (Nvidia RTX4070 12G 以上較佳)。 較易是用網上的平台,例如 tensor.art 去玩 SD,免費便會有限制。
POE/Chatgpt:
https://poe.com/
https://openai.com/
(Chatgpt need vpn, can use proton vpn:
https://protonvpn.com/download )
Bing (free to get image):
https://www.bing.com/images/create
安裝 Stable Diffusion (SD) webui:
https://github.com/AUTOMATIC1111/stable-diffusion-webui
較深入玩法,是利用工作流 comfyUI 去達成目的,
先安裝 comfyUI (已包含 python), manager 和語言 translator
ComfyUI install:
https://github.com/comfyanonymous/ComfyUI
裝一個 Manager 可找出 missing 節點,再由Manager自動安裝。
ComfyUI manager:
https://github.com/ltdrdata/ComfyUI-Manager
語言 translation:
https://github.com/AIGODLIKE/AIGODLIKE-ComfyUI-Translation
再裝, flux workflow v3 (Flux workflow):
https://www.stefanoangeli.it/comfyui-workflow-v3/
Flux 1.0:
https://huggingface.co/spaces/black-forest-labs/FLUX.1-schnell
https://huggingface.co/spaces/black-forest-labs/FLUX.1-dev
Tensor.art (free SD platform)
https://tensor.art/
use Midjourney (25 free photos):
https://www.midjourney.com/explore?tab=hot
Leonardo.ai
https://app.leonardo.ai/
Ideogram:
https://ideogram.ai/login
Krita to modify photo
https://krita.org/zh-tw/post-download/
Movie:
luma dream machine:
https://lumalabs.ai/dream-machine/creations
Sora: https://openai.com/index/sora/
影片生成,只用email 註册:
Kling國際版連結:https://klingai.com/
公仔/icon/virtual youtuber
Viggle AI:
https://viggle.ai/home
Gamma generate powerpoint
https://gamma.app/generate
Kimi ppt
https://kimi.moonshot.cn/kimiplus/conpg18t7lagbbsfqksg
Canva design:
https://www.canva.com/
Suno 歌曲: https://www.hedra.com/
song 廣東話拼音
https://hongkongvision.com/tool/cc_py_conv_zh
AI 自作歌曲(模擬人演釋): 廣東話歌曲 <情深似海>
https://www.facebook.com/inchinet/videos/389898824075586
Capcut Pro (China version) 再加字幕。