(有中文)Why Does AI Always Give You a Close-Up? How to Control Framing with a Four-Stage Ladder 📷✨

https://civarchive.com/articles/34198/why-does-ai-always-give-you-a-close-up-how-to-control-framing-with-a-four-stage-ladder

Once again, what I'm sharing today is an ultra-simple workflow. It follows my usual style: simple, practical, completely spaghetti-free, and 100% Low VRAM friendly.

All the info for the required nodes and LoRAs is already noted in the workflow, along with key things you need to watch out for.

For the full breakdown of the workflow, you can check out the article below. I won't repeat the details here:

用control net做出關鍵幀, 生成影片時就輕鬆多了

今次分享的亦是超簡單Workflow。

亦遵循我一貫作風, 簡單, 實際, 不會意粉化, Low VRAM能用

關於workflow的詳細可以參考以下文章, 這兒就不再重覆了

https://civarchive.com/articles/34090/how-to-set-a-complicated-pose-when-generating-a-video-in-minimax-h3

All credit belongs to the lora creator and node creator:👏

https://huggingface.co/Patil/Krea-2-depth-controlnet/tree/main

https://github.com/1038lab/ComfyUI-RMBG

https://github.com/facok/comfyui-krea2-controlnet