README.md

---
title: MiniMax H3 Prompt Enhancer
emoji: "🎬"
colorFrom: purple
colorTo: pink
sdk: gradio
sdk_version: 5.49.1
app_file: app.py
python_version: "3.10"
startup_duration_timeout: 1h
preload_from_hub:
  - geocine/minimax-video-prompt-enhancer-350m-gguf
pinned: false
license: other
short_description: MiniMax H3 video prompts via CPU GGUF
tags:
  - text-generation
  - video
  - prompt
  - minimax
  - liquid
  - lfm2
  - llama.cpp
  - cpu
---

# MiniMax H3 Prompt Enhancer

**Not another “make my prompt longer” toy.**

This Space runs a **purpose-trained small language model** that rewrites rough ideas into **production-ready MiniMax H3 video prompts** — the same kind of structured audiovisual brief MiniMax expects: shots, camera language, diegetic sound, and score, not a bag of adjectives.

Type something half-formed like *“baker opens a shop at dawn”*. The model doesn’t slap on “cinematic, 8k, masterpiece.” It **reasons** about scene, motion, and audio, then emits the **exact field layout** used for MiniMax video generation.

> **Want more creativity?** This Space runs the fast **350M** model. For a richer, more creative enhancer (style / camera / dialogue / music), open the
> [**2.6B creative Space**](https://huggingface.co/spaces/geocine/MiniMax-H3-Prompt-Enhancer-2.6B) (ZeroGPU).

---

## Why this is different

| Typical prompt “enhancers” | This one |
|---|---|
| Templates, regex, or a generic chat model | **Fine-tuned LLM** dedicated to this job |
| Mood words and hashtag soup | **Shot lists, camera moves, soundscape, music** |
| Guesswork about “what sounds good” | Aligned to **MiniMax H3 video prompt writing guidelines** |
| One free-form paragraph | **Structured output** MiniMax can actually consume |

Under the hood: LoRA SFT of [LiquidAI/LFM2.5-350M](https://huggingface.co/LiquidAI/LFM2.5-350M) —
transformers weights at [geocine/minimax-video-prompt-enhancer-350m](https://huggingface.co/geocine/minimax-video-prompt-enhancer-350m),
**Q4_K_M GGUF** at [geocine/minimax-video-prompt-enhancer-350m-gguf](https://huggingface.co/geocine/minimax-video-prompt-enhancer-350m-gguf),
served via **llama.cpp** on **CPU**.

---

## Runtime

| | |
|--|--|
| Backend | **llama.cpp** (`llama-cpp-python`) |
| Quant | **Q4_K_M** (~219 MB) |
| Hardware | **CPU** — no ZeroGPU quota, no login required for inference |
| Typical latency | A few seconds for a short T2VA brief on free CPU |

Weights are downloaded from the model repo at boot (or read from `models/` if present).

---

## What you get

Paste a rough idea (and optional reference notes). Choose duration. Hit **Generate**.

You get a full brief, for example:

- **`integrated_multimodal_description`** — style, framing, action, camera, dialogue tags  
- **`overall_soundscape`** — ambience and physical sound  
- **`non_diegetic_music`** — audience-only score (or N/A)  
- For image/video refs: alignment lines, subjects, retention-style structure in **full-reference** mode  

Built for people who actually ship MiniMax clips — not for prompt-cosplay.

---

## How to use

1. Write a rough user prompt (dialogue, shot notes, or a one-liner).  
2. Optionally fill **References & frames** (first/last frame, subjects, clips, audio). The Space picks the task for you.  
3. Set duration. Generate. Copy the structured prompt into MiniMax.

**Tip:** Short, real-user prompts work best. The model expands them into the full MiniMax format.

---

## Model & license

| | |
|--|--|
| Model (HF) | [geocine/minimax-video-prompt-enhancer-350m](https://huggingface.co/geocine/minimax-video-prompt-enhancer-350m) |
| GGUF (CPU) | [geocine/minimax-video-prompt-enhancer-350m-gguf](https://huggingface.co/geocine/minimax-video-prompt-enhancer-350m-gguf) |
| Base | [LiquidAI/LFM2.5-350M](https://huggingface.co/LiquidAI/LFM2.5-350M) |
| License | Liquid **lfm1.0** — read the base card before commercial use |

This is a **text prompt specialist**, not a video generator. It writes the brief; MiniMax (or your pipeline) makes the frames.