MiniMax-H3-Image-FP8
unsloth/MiniMax-H3-Image-FP8
- 作者
- unsloth
- 任务
- 图像生成
- 下载量
- 2,965
- 点赞数
- 9
- 趋势分
- 8
- 许可证
- other
- 收录时间
- 2026/11/13
base_model:
- MiniMaxAI/MiniMax-H3-Image base_model_relation: quantized tags:
- fp8
- int8
- unsloth
- quantized
- minimax-h3
- image-generation
- image-editing
- text-to-image license: other license_name: minimax-h3-image-community-license-agreement license_link: https://huggingface.co/MiniMaxAI/MiniMax-H3-Image/blob/main/LICENSE language:
- en pipeline_tag: text-to-image widget:
- text: "A weathered lighthouse keeper in a yellow oilcloth coat stands on a wet stone pier at dusk, holding a brass lantern whose warm light falls across his lined face and grey beard. Behind him a white lighthouse with a red band rises against a stormy violet sky, its beam cutting through drifting sea spray. Waves break against black rocks on the left, and a small fishing boat with a green hull is tied to an iron ring in the foreground. Cinematic photograph, shallow depth of field, rich detail in the fabric, the stone and the water." output: url: samples/int8_1.png
- text: "A cozy independent bookshop interior seen from the doorway on a rainy afternoon. Floor-to-ceiling wooden shelves packed with colourful books line both walls, a rolling library ladder leans against the right shelf, and a ginger cat sleeps on a stack of hardcovers on a round oak table in the centre. A hand-lettered chalkboard sign on the table reads "NEW ARRIVALS" in white chalk. Warm pendant lamps glow overhead, rain streaks the front window, and a green armchair with a knitted blanket sits in the back corner. Natural, detailed, photographic." output: url: samples/fp8_2.png
- text: "A detailed isometric illustration of a tiny floating island city at sunrise. Terraced pastel houses with terracotta roofs climb a rocky hill crowned by a glass greenhouse dome, a waterfall pours off the island's edge into soft clouds below, and small hot-air balloons in red, teal and mustard drift around it. Winding stairways, lantern-lit bridges and tiny people walking dogs fill the streets. Clean lines, soft gradients, crisp shadows, a calm pastel palette." output: url: samples/int8_3.png
Run MiniMax-H3-Image with Unsloth 💜
This is an INT8 / FP8 quantized version of the MiniMax-H3-Image diffusion transformer.
- Runs on a 24 GB L4 with the DiT streamed from RAM (19.4 GB peak VRAM, about 50 s per 2048x2048 image in FP8), or on a 40 GB A100 with the INT8 DiT resident (about 31 s).
- Same-seed LPIPS vs the bf16 model at 2048x2048 (lower is better, mean of 4 prompts): 0.140 INT8-ConvRot, 0.255 FP8. INT8-ConvRot is the recommended scheme; FP8 quantized on the fly scores 0.266, so its gap is FP8 itself, not this file.
- In Unsloth Desktop these are used automatically: pick MiniMax-H3-Image and an INT8 or FP8 precision.
| file | size | replaces |
|---|---|---|
MiniMax-H3-Image-INT8-ConvRot.safetensors | 20.27 GB | the 66.3 GB bf16 transformer and its per-step heads |
MiniMax-H3-Image-FP8.safetensors | 20.27 GB | the 66.3 GB bf16 transformer and its per-step heads |
The VAE, tokenizer and schedule come from MiniMaxAI/MiniMax-H3-Image, and the text encoder from MiniMaxAI/MiniMax-H3 (an INT8 copy is in unsloth/MiniMax-H3-FP8). Keep context_embedder and token_refiner in bf16 at load (Unsloth does): with them quantized, LPIPS rises to 0.161 INT8 / 0.308 FP8. Everything here is safetensors rather than a pickle, so loading runs no arbitrary code. These files are Model Derivatives: the DiT is quantized and its modulation stored in curve form, as listed in NOTICE.
Samples
MiniMax-H3-Image
Online API
Use MiniMax-H3-Image directly via API.
- Global: platform.minimax.io | CN: platform.minimax.cn
API docs:
Online App
Use MiniMax-H3-Image directly via App.
- WebApp Global: hailuoai.video | CN: hailuoai.com
- Desktop Global: agent.minimax.io | CN: agent.minimax.cn
Introduction
MiniMax-H3-Image is a state-of-the-art open-source model for image generation and editing from the MiniMax H3 family. It combines intelligent understanding of creative intent with the ability to build visual worlds: with precise instruction following, it develops a creative concept into concrete characters, scenes and atmosphere. Generation and editing are unified in one checkpoint, which generates images from text (text-to-image) and edits one or more reference images (image editing), giving creators and developers a flexible and controllable way to create images.
It generates images of about 16 megapixels at every aspect ratio, from 4096x4096 (1:1) to 6272x2688 (21:9), directly and without an upscaler (see Sizes). The checkpoint is distilled to 6 sampling steps (6 NFEs) with no classifier-free guidance. Its diffusion transformer (DiT) has the H3 architecture without the audio layers, and it uses the MiniMax-H3 text encoder unchanged.
Highlights
MiniMax-H3-Image brings together three core strengths: speed, image quality, and instruction following.
- 16-megapixel images at a practical computational cost. Compared with the 1–4 MP outputs common today, 16 MP (4096 × 4096) images offer far more detail but typically require substantially more computation. MiniMax-H3-Image keeps that computational cost manageable, making 16 MP generation practical for everyday creation.
- Rich detail. Remarkable realism. MiniMax-H3-Image goes beyond the familiar “AI look” to produce images with intricate detail and lifelike realism, opening up more possibilities for creative and professional use. It brings generated imagery closer to the richness and nuance of the real visual world.
- Stronger instruction following. Greater creative control. Great image generation goes beyond beautiful pictures—it helps creators bring their ideas to life. MiniMax-H3-Image delivers the strongest instruction following among open-source models and supports up to 16 reference images, giving creators more control over the final result.
Model Zoo
| Model | Role | Precision | Size | Hugging Face |
|---|---|---|---|---|
| MiniMax-H3-Image | Text-to-image and image editing | BF16 (DiT) | 68.65 GB | MiniMaxAI/MiniMax-H3-Image |
| MiniMax-H3-Image-PE (optional) | Prompt expander for text-to-image | NVFP4 | 17.9 GB | MiniMaxAI/MiniMax-H3-Image-PE |
Model Architecture
MiniMax-H3-Image follows the H3-Base design with the audio branch removed. The text prompt and the reference images are encoded by the H3 text encoder; the reference images are also encoded by the image VAE. All tokens form one packed sequence for a single-stream transformer, which predicts the latent of the target image. The VAE decodes that latent into the image.
- Text encoder: the full Qwen3-VL text encoder of MiniMax-H3, unchanged. The DiT reads the hidden states of its 50th layer.
- Image VAE: a CNN encoder and a 2D ViT decoder (24 layers), with 16x spatial downsampling and 24 latent channels. The DiT patchifies the latent with a 2x2 patch, so each transformer token covers 32x32 pixels.
- Diffusion transformer: 33.1B parameters, a dense single-stream transformer with 50 layers (hidden size 5376, 56 attention heads, FFN size 14336) and 3D multimodal RoPE, the same as the H3-Omni-Transformer. The weights are BF16; the embedding, timestep and output layers are FP32.
Quick Start
Download
| Component | Content | Size |
|---|---|---|
transformer/ | DiT: BF16, with the embedding, timestep and output layers in FP32. | 66.28 GB |
pdd_fused_heads.safetensors, pdd_config.json | The per-step output heads that replace the transformer's output layer, and the sampling schedule. | 12 MB |
vae/ | Image VAE, FP32. | 2.34 GB |
tokenizer/, processor/ | Qwen3-VL tokenizer and processor, identical to MiniMax-H3. | 23 MB |
| Text encoder | Not in this repository. model_index.json loads it from MiniMaxAI/MiniMax-H3, folder text_encoder/. If you already have MiniMax-H3, it is not downloaded again. | 66.7 GB |
MiniMax-H3-Image/
├── model_index.json
├── pdd_config.json
├── pdd_fused_heads.safetensors
├── processor/
├── tokenizer/
├── transformer/
└── vae/
Download this repository and the MiniMax-H3 text encoder into the Hugging Face cache:
hf download MiniMaxAI/MiniMax-H3-Image
hf download MiniMaxAI/MiniMax-H3 --include "text_encoder/*"
# optional prompt expander
hf download MiniMaxAI/MiniMax-H3-Image-PE
Local Deployment
We recommend the following inference frameworks to serve the model:
- SGLang - see cookbook
- vLLM-Omni - see vllm-omni recipe
- diffusers - see diffusers docs
- ComfyUI - see Comfy tutorial; use T2I template / Edit template
vLLM-Omni Deployment
See the vLLM-Omni recipe for per-hardware deployment configurations and measured latency. Both NVIDIA and AMD accelerators are supported, including NVIDIA H200, RTX PRO 6000 Blackwell, AMD Instinct MI350X and more.
Start the server on one GPU:
vllm serve MiniMaxAI/MiniMax-H3-Image --omni --host 0.0.0.0 --port 8000
For lower latency, split each request over several GPUs with sequence parallelism (Ulysses), for example on 4 GPUs:
vllm serve MiniMaxAI/MiniMax-H3-Image --omni --host 0.0.0.0 --port 8000 \
--num-gpus 4 --usp 4 --text-encoder-tp-size 4
SGLang Deployment
Here we use SGLang as a deployment example. See the MiniMax-H3-Image deployment guide for additional deployment configurations.
Start the server on one GPU:
sglang serve --model-path MiniMaxAI/MiniMax-H3-Image --host 0.0.0.0 --port 30000 --warmup-mode off
To use a local copy of the MiniMax-H3 text_encoder/ folder instead of the Hugging Face
cache, add --text-encoder-path <path to text_encoder>.
For lower latency, split each request over several GPUs with sequence parallelism (Ulysses), for example on 4 GPUs:
sglang serve --model-path MiniMaxAI/MiniMax-H3-Image --host 0.0.0.0 --port 30000 --warmup-mode off \
--num-gpus 4 --ulysses-degree 4
Request parameters
Text-to-image requests are JSON sent to /v1/images/generations. Image editing requests send
the reference images as multipart form data to /v1/images/edits; their order defines
<Picture 1>, <Picture 2>, and so on in the prompt. Both return the PNG in
data[0].b64_json. Runnable examples are in Full Workflow.
| Field | Text-to-image | Image editing | Notes |
|---|---|---|---|
prompt | Required | Required | Used as given, truncated to 4096 tokens. See Full Workflow |
size | Required | Required | WIDTHxHEIGHT. The returned image has exactly this size; see Sizes |
image | — | One or more files | In upload order they are <Picture 1>, <Picture 2>, ... in the prompt |
seed | Recommended | Recommended | Integer |
generator_device | cpu recommended | cpu recommended | Same seed, same initial noise on every machine |
response_format | b64_json | b64_json | |
min_generation_area | Optional | Optional | Minimum generated area in pixels, 1024-16777216, default 4194304 (2048x2048). A smaller size is generated at the same aspect ratio on a larger canvas and resized back with LANCZOS; pass 1024 to turn this off. All sizes in the table below are above the default |
num_inference_steps | Omit or 6 | Omit or 6 | Other values return HTTP 400 |
guidance_scale | Omit or 1 | Omit or 1 | Other values return HTTP 400 |
negative_prompt | Not supported | Not supported | Rejected |
Sizes
The effective working range of MiniMax-H3-Image is about 4 MP to 20 MP, and it generally gives the best results at around 16 MP. Common 16 MP sizes are listed below:
| Aspect ratio | Size (width x height) |
|---|---|
| 1:1 | 4096x4096 |
| 4:3 / 3:4 | 4736x3552 / 3552x4736 |
| 3:2 / 2:3 | 5024x3360 / 3360x5024 |
| 16:9 / 9:16 | 5472x3072 / 3072x5472 |
| 21:9 | 6272x2688 |
Full Workflow
MiniMax-H3-Image expects a detailed English description of the whole image. The MiniMax API turns a short prompt in any language into such a description with H3-Image-Context-IR before generation. We also provide MiniMax-H3-Image-PE, which can be deployed locally.
This section reproduces the quality of images generated directly by the MiniMax API with a
locally deployed SGLang server. The scripts are in
examples/ and use only
the Python standard library; run them in that folder:
hf download MiniMaxAI/MiniMax-H3-Image --include "examples/*" --local-dir .
cd examples
Before running examples, set your MiniMax API endpoint and key to use the H3-Image-Context-IR API and the reference scripts. The scripts read them from the environment:
# MiniMax API endpoint (choose one)
# CN
export MINIMAX_API_BASE="https://api.minimax.cn"
# Global
# export MINIMAX_API_BASE="https://api.minimax.io"
# API key obtained from the MiniMax platform
export MINIMAX_API_KEY="<api-key>"
API docs:
- H3-Image-Context-IR: use /v1/h3_image_context_ir EN-docs, CN-docs
- Text-to-image: use /v1/images/generations EN-docs, CN-docs
- Image editing: use /v1/images/edits EN-docs, CN-docs
case-T2I
case-Edit
instruction.edit.txt
with the reference image reference1.png
as <Picture 1>.
| Stage | Request | Result |
|---|---|---|
| Prompt Enhancement: H3-Image-Context-IR (MiniMax API) | python3 context_ir.py edit (script) | revised_prompt.txt, revised_size.txt; our run: revised_prompt.edit.txt, revised_size.edit.txt (4096x4096) |
| MiniMax-H3-Image generation: local SGLang | python3 edit.py --prompt revised_prompt.txt --size $(cat revised_size.txt) (script) | edit.png |
| E2E API Reference: image editing by directly calling the MiniMax API | python3 api_generate.py edit (script) | api_edit.png |
Showcase
Text-to-image and image editing results, shown downscaled.
License
- MiniMax-H3-Image is released under the MiniMax-H3-Image Community License Agreement.
- The T2I PE Model (MiniMax-H3-Image-PE) is part of MiniMax-H3-Image and is accordingly subject to MiniMax-H3-Image Community License Agreement.
- The separate written authorization described in Section IV of the license is requested through Application form
Contact Us
Contact us at model@minimax.io.
内容来源：Hugging Face 模型卡
在 Hugging Face 查看