---
title: minimax_h3_fl2v_lightx2v_4step_int8-convrot_comfy
canonical_url: "https://www.modelscope.cn/models/rzgar/minimax_h3_fl2v_lightx2v_4step_int8-convrot_comfy"
md_url: "https://www.modelscope.cn/models/rzgar/minimax_h3_fl2v_lightx2v_4step_int8-convrot_comfy.md"
repository: rzgar/minimax_h3_fl2v_lightx2v_4step_int8-convrot_comfy
last_updated: 2026-08-13
license: apache-2.0
pipeline_tag: text-to-video-synthesis
tasks:
  - text-to-video-synthesis
base_model:
  - MiniMaxAI/MiniMax-H3
base_model_relation: adapter
parameters: 8.7B
tensor_type:
  - BF16
  - U8
  - I8
  - F32
library_name:
  - lora
  - diffusers
  - peft
  - safetensors
  - pytorch
frameworks:
  - pytorch
downloads: 1924
stars: 8
tags:
  - text-to-video
  - text-to-audio
  - audio-video
  - lora
  - minimax-h3
  - diffusers
  - peft
---

# minimax_h3_fl2v_lightx2v_4step_int8-convrot_comfy

> minimax_h3_fl2v_lightx2v_4step_int8-convrot_comfy - rzgar 在 ModelScope 开源的模型。MiniMax H3 FL2V LightX2V 4-Step Int8-ConvRot

rzgar/minimax_h3_fl2v_lightx2v_4step_int8-convrot_comfy 是 ModelScope 魔搭社区上的 8.7B 参数text-to-video-synthesis模型，采用 apache-2.0 许可，基于 MiniMaxAI/MiniMax-H3 构建。

- **Repository**: rzgar/minimax_h3_fl2v_lightx2v_4step_int8-convrot_comfy
- **License**: apache-2.0
- **Tasks**: text-to-video-synthesis
- **Parameters**: 8.7B
- **Base model**: MiniMaxAI/MiniMax-H3
- **Tags**: text-to-video, text-to-audio, audio-video, lora, minimax-h3, diffusers, peft
- **Downloads**: 1924
- **Stars**: 8
- **Last updated**: 2026-08-13

Source: https://www.modelscope.cn/models/rzgar/minimax_h3_fl2v_lightx2v_4step_int8-convrot_comfy

---

# MiniMax H3 FL2V LightX2V 4-Step Int8-ConvRot 


_Clip 1: v1 4step | Clip 2: v1 8step_
<video controls width="1280" height="720">
  <source src="https://huggingface.co/rzgar/minimax_h3_fl2v_lightx2v_4step_int8-convrot_comfy/resolve/main/demo/demo_vl_4_8step.mp4" type="video/mp4">
  Your browser does not support the video tag.
</video>

_v0.1 4step demo_
<video controls width="1280" height="720">
  <source src="https://huggingface.co/rzgar/minimax_h3_fl2v_lightx2v_4step_int8-convrot_comfy/resolve/main/demo/MiniMax_H3_00108_.mp4" type="video/mp4">
  Your browser does not support the video tag.
</video>

<Gallery />

## Model description 


## Requirements

- ComfyUI (recent - tested on master 0.31.0)
- **[ComfyUI-LoraInt8Loader](https://huggingface.co/rzgar/minimax_h3_fl2v_lightx2v_4step_int8-convrot_comfy/resolve/main/ComfyUI-LoraInt8Loader/ComfyUI-LoraInt8Loader.zip)** custom node (standard ComfyUI LoRA loaders cannot dequantize it)

## Files

| File | Size | NDownload |
|---|---|---|
| `minimax_h3_fl2v_turbo_4step_v1.0_int8_convrot` | 991 MB | [download](https://huggingface.co/rzgar/minimax_h3_fl2v_lightx2v_4step_int8-convrot_comfy/resolve/main/minimax_h3_fl2v_turbo_4step_v1.0_768p_comfyui_bf16_int8convrot.safetensors)|
| `minimax_h3_fl2v_turbo_8step_v1.0_int8_convrot` | 991 MB | [Download](https://huggingface.co/rzgar/minimax_h3_fl2v_lightx2v_4step_int8-convrot_comfy/resolve/main/minimax_h3_fl2v_turbo_8step_v1.0_comfyui_bf16_int8convrot.safetensors) |
| `minimax_h3_fl2v_turbo_4step_v0.1_int8_convrot` | 991 MB | [Download](https://huggingface.co/rzgar/minimax_h3_fl2v_lightx2v_4step_int8-convrot_comfy/resolve/main/minimax_h3_fl2v_lightx2v_turbo_4step_v0.1_comfy_int8convrot.safetensors) | 0.99 GB | Full LoRA |

- Both are quantized from the [KJ comfy-converted BF16 LoRA](https:&#x2F;&#x2F;huggingface.co&#x2F;Kijai&#x2F;MiniMax-H3_comfy&#x2F;tree&#x2F;main&#x2F;loras).

- [LightX2V Release](https://huggingface.co/lightx2v/Minimax-h3-Turbo/tree/main)

## norefiner variant

The full LoRA also patches the model's token_refiner - the text/audio
conditioning refiner. The norefiner variant removes those 8 patches so the
conditioning path runs at base quality (_some video clips had bad audio. This solution fixed it, but keep in mind it doesn’t happen always_). 
If audio/text conditioning feels weak with the full LoRA at 4 steps, try the norefiner variant.

_First clip: with refiner | last clip: norefiner_
<video controls width="1280" height="720">
  <source src="https://huggingface.co/rzgar/minimax_h3_fl2v_lightx2v_4step_int8-convrot_comfy/resolve/main/demo/demo.mp4" type="video/mp4">
  Your browser does not support the video tag.
</video>


## Usage (ComfyUI)

1. Unzip the **ComfyUI-LoraInt8Loader** custom node into **custom_nodes**, restart
2. In your MiniMax H3 workflow, add **LoRA Loader (Int8-ConvRot)**, select the
   file, **strength 1.0** (the training scale is already baked in)
3. Sample at 4-6 steps


## Download model


[Download](/rzgar/minimax_h3_fl2v_lightx2v_4step_int8-convrot_comfy/tree/main) them in the Files & versions tab.