Resources

【招聘】 MiniMax 视频模型全球开源生态

pinned

❤️ 4

#61 opened 15 days ago by

MiniMax-AI

Any License Question Ask here!

pinned

👍 16

#12 opened 19 days ago by

ryanlee-dev

Make new

#94 opened 1 day ago by

hkb4181

Running on <4gb Vram

🔥 1

#93 opened 1 day ago by

Jit2024

H3-Regenerate-2K open weights?

➕ 1

#92 opened 3 days ago by deleted

MiniMax H3 Ref2VA + audio recasts identity, while FL2VA + Ref2V Turbo preserves it — expected behavior?

#91 opened 3 days ago by

lorentaken

PURE MOJO STACK

#90 opened 3 days ago by

alexnvo

蒸馏加速版本

#89 opened 3 days ago by

ReSerendipity

Any plans to release a non-distilled base model for training?

👍 2

#88 opened 4 days ago by

AndroYD

脸崩的如此厉害，何时修复~

#87 opened 4 days ago by

criuslv-cn

This is currently the best open‑source video model, yet the conservative algorithm of the VAE seems to create a computational bottleneck?

👍❤️ 8

#85 opened 6 days ago by

ass5002

Getting different Accents when speaking the same language

#84 opened 6 days ago by

Hcheyne

Ref2VA environmental audio feels too quiet in dialogue scenes

➕ 1

#82 opened 6 days ago by

szczypen

Question about prompting on static elements

#81 opened 7 days ago by

leonnn1

Thanks and Suggestions

#80 opened 8 days ago by

Pvok

Omni-Rewriter Replay: observe a clip into validated H3 PE (open harness)

#79 opened 8 days ago by

Wayne-King

Color distortion occurs when encoding and decoding images using a video VAE.

#78 opened 9 days ago by

l13462580123

checkpoint inconsistent between original and diffusers

#77 opened 9 days ago by

l13462580123

Weird gibberish / clipped voice at the very start of H3 videos (reference to video)

👍 4

#76 opened 9 days ago by

jesleocizi

[FL2VA] Anime style videos have half the FPS

#74 opened 10 days ago by

ChiNoel

Call MiniMax H3 with your existing OpenAI SDK

#72 opened 10 days ago by

irene891107

Some question about VIDEO_PROMPT_WRITING_GUIDE_ref_en.md

#71 opened 12 days ago by

iouzzr

Minimax H3 prompt adherence really varies depending on the resolution

#65 opened 14 days ago by

TheBobun

How define audio ref to the Subject 1 - Ref2V

#64 opened 14 days ago by

Gamb

第一次接触咱们这个网站，有好几个问题想要问

#63 opened 14 days ago by

zhao7259

Update documentation to Please add how to character swap for Ref2V/Image to Video.

#60 opened 15 days ago by

CoolaidFun

脱领带，脱上衣，脱带搭袢的鞋等还无法正常表现

#59 opened 15 days ago by

sumirecccp

MiniMax H3 Prompt Enhancer, powered by a fine-tuned 350M-parameter model

❤️ 3

#58 opened 15 days ago by

geocine

Will RunPod work??

#56 opened 16 days ago by

tonyface

Almost VR support

👀 3

#55 opened 16 days ago by

Ddfgddsd

请教一下，这个模型如何生成一个可以首尾循环的视频？

#54 opened 16 days ago by

oioitff

lora训练有什么需要注意的地方？

#53 opened 16 days ago by

wuyuetiger

Ref2va always have some noise

👀❤️ 2

#50 opened 17 days ago by

jiangjihua

Vllmomni minmaxh3 roadmap

#49 opened 17 days ago by

feizhai123

sparse-attention inference

#48 opened 17 days ago by

mzbac

The actual correct system prompt for IT2V - (Took me a while)

👍 6

#47 opened 17 days ago by

cushycrux

Where is the money, Lebowski? The "duct-tape" architecture review)))

➕ 3

#46 opened 17 days ago by

Qozimo

Possibly a small mistake in docs/VIDEO_PROMPT_WRITING_GUIDE_base_en.md

#45 opened 17 days ago by

hum-ma

Is this a CFG distilled model or were the weights trained without CFG from the start?

👍👀 3

#44 opened 18 days ago by

natalie5

LoRA Training with MiniMax H3 - Question

#43 opened 18 days ago by

Tomcat2048

Thanks a lot

❤️ 14

#42 opened 18 days ago by

DodoPapa

我是阿里的高管

😎 5

#41 opened 18 days ago by

wjm17173

Amazing job ! really thank you for sharing and pushing forward the video generation technology!

👍 1

#40 opened 18 days ago by

aiclouddaily

Minimax H3 2K Upscaler?

#39 opened 18 days ago by

AlperKTS

Only MiniMax-H3 can do !

🔥 5

#38 opened 18 days ago by

sunnyboxs

非常棒的多模态能力。是否可以基于这个扩展生图能力？

#37 opened 18 days ago by

wdtd

vLLM-Omni ComfyUI integration for MiniMax-H3 (T2VA / FL2VA / Ref2VA)

🚀 2

#36 opened 18 days ago by

shunyang90

my opinion about nsfw

😎 14

#35 opened 18 days ago by

Arun63

Is there any plan to release a more lightweight model with fewer parameters?

➕ 2

#34 opened 18 days ago by

makisekurisu-jp

请问支持中文提示词吗？相关的特殊tag需要替换成中文吗？

#32 opened 18 days ago by

HappyNeNe