Resources

【招聘】 MiniMax 视频模型全球开源生态

pinned

❤️ 4

#61 opened 9 days ago by

MiniMax-AI

Any License Question Ask here!

pinned

👍 16

#12 opened 14 days ago by

ryanlee-dev

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

❤️ 1

#85 opened about 7 hours ago by

ass5002

Getting different Accents when speaking the same language

#84 opened about 7 hours ago by

Hcheyne

Very bad prompt adherence

#83 opened about 10 hours ago by

gemstonebro

Ref2VA environmental audio feels too quiet in dialogue scenes

#82 opened about 17 hours ago by

szczypen

Question about prompting on static elements

#81 opened 2 days ago by

leonnn1

Thanks and Suggestions

#80 opened 3 days ago by

Pvok

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

#79 opened 3 days ago by

Wayne-King

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

#78 opened 4 days ago by

l13462580123

checkpoint inconsistent between original and diffusers

#77 opened 4 days ago by

l13462580123

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

👍 2

#76 opened 4 days ago by

jesleocizi

[FL2VA] Anime style videos have half the FPS

#74 opened 4 days ago by

ChiNoel

Call MiniMax H3 with your existing OpenAI SDK

#72 opened 5 days ago by

irene891107

Some question about VIDEO_PROMPT_WRITING_GUIDE_ref_en.md

#71 opened 6 days ago by

iouzzr

Minimax H3 prompt adherence really varies depending on the resolution

#65 opened 8 days ago by

TheBobun

How define audio ref to the Subject 1 - Ref2V

#64 opened 8 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 8 days ago by

zhao7259

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

#60 opened 9 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 9 days ago by

sumirecccp

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

❤️ 3

#58 opened 10 days ago by

geocine

Will RunPod work??

#56 opened 10 days ago by

tonyface

Almost VR support

👀 3

#55 opened 10 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 10 days ago by

oioitff

lora训练有什么需要注意的地方？

#53 opened 10 days ago by

wuyuetiger

Ref2va always have some noise

👀❤️ 2

#50 opened 11 days ago by

jiangjihua

Vllmomni minmaxh3 roadmap

#49 opened 12 days ago by

feizhai123

sparse-attention inference

#48 opened 12 days ago by

mzbac

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

👍 5

#47 opened 12 days ago by

cushycrux

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

➕ 3

#46 opened 12 days ago by

Qozimo

Possibly a small mistake in docs/VIDEO_PROMPT_WRITING_GUIDE_base_en.md

#45 opened 12 days ago by

hum-ma

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

👍👀 3

#44 opened 12 days ago by

natalie5

LoRA Training with MiniMax H3 - Question

#43 opened 12 days ago by

Tomcat2048

Thanks a lot

❤️ 13

#42 opened 12 days ago by

DodoPapa

我是阿里的高管

😎 5

#41 opened 12 days ago by

wjm17173

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

👍 1

#40 opened 12 days ago by

aiclouddaily

Minimax H3 2K Upscaler?

#39 opened 12 days ago by

AlperKTS

Only MiniMax-H3 can do !

🔥 5

#38 opened 12 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 12 days ago by

wdtd

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

🚀 1

#36 opened 12 days ago by

shunyang90

my opinion about nsfw

😎 14

#35 opened 13 days ago by

Arun63

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

➕ 2

#34 opened 13 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 13 days ago by

HappyNeNe

简直是圣人

#31 opened 13 days ago by

jyz1022

how much better is the cloud version of h3?

#29 opened 13 days ago by

BrokeAmerican

System Prompt IT2V

👀🔥 17

#28 opened 13 days ago by

rzgar

No official trainer? We open-sourced a working fine-tuning pipeline for H3

🚀 4

#27 opened 13 days ago by

ka1029

Go minimax!, fxxx seedance for ridiculous pricing.

🚀 15

#26 opened 13 days ago by

anitman

Future request

👍 2

#25 opened 13 days ago by

darkenHUB

Huge Thanks for Using a Single-Stream Transformer, Porting It to a Multi-GPU ComfyUI Setup with RayLight Is a Breeze

#24 opened 13 days ago by

komixenon