Low VRAM專用簡單Minimax H3 Prompt Enhancer Workflow
Easy Prompt Enhancer Workflow for Minimax H3 (Low VRAM)
The purpose of separating the Prompt Enhancer from the rest of the workflow is to reduce the VRAM load.
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.
I’ve covered the rest of the details directly inside the workflow and the article below, so I won't repeat them here.
Aside from saving on token consumption, the primary goal of using Local LLMs is to bypass the heavy censorship found in closed-source models. Closed-source LLMs are all filtered and simply won't respond to NSFW requests. For example, I wanted to generate an image of a girl holding a gun and shooting at the camera, but closed-source models are incapable of that. Forget about the more sensitive topics like pornography or violence. 🚫🔫
把Prompt Enhancer與生成短片的Workflow分開目的，就是為了減輕VRAM負擔，提供用家選項, 令介面更簡潔乾淨。
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
其他詳細在workflow和下方文章中有說明就不再重覆了。
用Local LLM除了省token消費外, 最重心的目的就是閉源的LLM全都是有過濾, NSFW的東西根本不會回答, 像本來我想生成的, 是一個少女拿著手槍向鏡頭開槍, 閉源的就是生成不了, 更不用說色情暴力一類。
我把上方這句中文傳給Gemini後換來以下的回答
and this is my Local LLM's answer:
然後這是Local LLM的答案:
(有中文)System prompt of Local LLM to help write prompts for MiniMax H3
How to use local LLM to write prompt for Minimax H3 (Low VRAM)
https://civarchive.com/articles/34255/how-to-use-local-llm-to-write-prompt-for-minimax-h3-low-vram
How to use ChatGPT/Gemini/Grok to help write prompts for MiniMax H3?
Special Thanks:
Video Upscale simple workflow for low VRAM (Extreme FAST)
Low VRAM專用簡單影片放大Workflow (極快)
https://civarchive.com/models/2871874/video-upscale-workflow-for-low-vram-extreme-fast
How to merge videos with FFmpeg in ComfyUI Desktop (2026)?
如何使用FFmpeg在ComfyUi Desktop(2026)中合拼片段
https://civarchive.com/articles/34119/how-to-merge-videos-with-ffmpeg-in-comfyui-desktop-2026
This is a simple workflow for VoxCPM2, featuring a single-speaker setup and a multi-speaker setup.
這是作為輸助語音生成VoxCPM2的Workflow
https://civarchive.com/models/2865730/voxcpm2-simple-workflow-for-low-v-ram
I've also written an article about VoxCPM2—feel free to check it out if you're interested!
還有關於VoxCPM2的文章, 有興趣可以看看
Krea 2 Quadview Simple workflow
Qaudview might lock the character appearence during mulit reference in long video generation.
Krea 2 多角度參考圖 Workflow
多角度參考圖可以在影片生成時鎖死角色外表, 特別在長片分多次生成中作為參考圖使用
https://civarchive.com/models/2867878/krea-2-quadview-simple-workflow-for-low-vram
Description
make sure you had choose the right LLM model and place it in comfyui/models/LLM
and choose smaller gguf while you are LOW VRAM user
most of the download link had mention in the workflow, read it before use it!
請確保您已選擇正確的 LLM 模型並將其放置在 comfyui/models/LLM 目錄中，如果您是低顯存用戶，請選擇較小的 gguf 檔案。大多數下載連結在工作流程中都有提及，請在使用前閱讀！