Image-Text-to-Video
Diffusers
Safetensors
MiniMaxH3ModularPipeline
sdnq
uint4
quantized
text-to-audio-video
image-to-audio-video
reference-to-audio-video
8-bit precision
Instructions to use ai-server-mgr/MiniMax-H3-SDNQ-uint4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use ai-server-mgr/MiniMax-H3-SDNQ-uint4 with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("ai-server-mgr/MiniMax-H3-SDNQ-uint4", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
Download NOTICE from ai-server-mgr/MiniMax-H3-SDNQ-uint4: direct link, hf CLI and curl.
- Browser
- Download file 777 Bytes
-
https://huggingface.co/ai-server-mgr/MiniMax-H3-SDNQ-uint4/resolve/main/NOTICE
- Command line
-
hf download hf://ai-server-mgr/MiniMax-H3-SDNQ-uint4/NOTICE
-
curl -L -o NOTICE https://huggingface.co/ai-server-mgr/MiniMax-H3-SDNQ-uint4/resolve/main/NOTICE
777 Bytes
| MiniMax H3 is licensed under the MiniMax H3 Community License Agreement, Copyright © 2026 MiniMax. All Rights Reserved. | |
| This repository contains an independently quantized derivative of MiniMaxAI/MiniMax-H3. | |
| Source revision: 42ed227ee7df40d41602854ae760620d6eb651fe. | |
| Modified weights: transformer, transformer_ref, and text_encoder, converted to SDNQ UINT4 with group size 64. | |
| Modified configurations: SDNQ quantization/shard metadata and portable pipeline indices. | |
| The model card and validation report describe the independent conversion and its limited validation. | |
| The MiniMax H3 encoder uses Qwen3-VL-32B, licensed under Apache License 2.0: | |
| https://github.com/QwenLM/Qwen3-VL/blob/main/LICENSE | |
| No change to upstream model licensing or endorsement by MiniMax is implied. | |