humanaigc/emo

Emote Portrait Alive: Generating Expressive Portrait Videos with Audio2Video Diffusion Model under Weak Conditions

EMO

Emote Portrait Alive: Generating Expressive Portrait Videos with Audio2Video Diffusion Model under Weak Conditions

Linrui Tian, Qi Wang, Bang Zhang, Liefeng Bo,

Institute for Intelligent Computing, Alibaba Group

Citation

@misc{tian2024emo,
      title={EMO: Emote Portrait Alive - Generating Expressive Portrait Videos with Audio2Video Diffusion Model under Weak Conditions}, 
      author={Linrui Tian and Qi Wang and Bang Zhang and Liefeng Bo},
      year={2024},
      eprint={2402.17485},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}

emo's Issues

Fraud in the AI industry, Vaporware aka "Proof of concept, or GTFO"

HumanAIGC/AnimateAnyone#27

Also, python devs - please remember to include the python version in your documentation!

[Non-Official Updates] Follow-up plans for the project

Thank you all for your incredible support and interest in our project. We've received lots of inquiries regarding a demo or the source code. We want to assure you that we are actively working on preparing the demo and code for public release. Although we cannot commit to a specific release date at this very moment, please be certain that the intention to provide access to both the demo and our source code is firm.

Our goal is to not only share the code but also ensure that it is robust and user-friendly, transitioning it from an academic prototype to a more polished version that provides a seamless experience. We appreciate your patience as we take the necessary steps to clean, document, and test the code to meet these standards.

Thank you for your understanding and continuous support.

HumanAIGC/AnimateAnyone#12

The chance of open source of this project is 1/6, here is my PROOF:

发尼玛啊，**开源搞不起来的风气就是被阿里这种B企业带坏的，阿里一生黑

继续骗赞吧

开源么

一贯作风，改参数、堆数据发个论文，再花钱宣传一波，最后再刷一下star混混kpi

为什么OpenAI的Sora不开源没有人说，而阿里巴巴的不开源却被骂？

双标的真是可以。

搞个空项目，你说个J8

J8J8J8J8J8J8J8J8J8J8J8J8J8J8

Where's the code?

Why post this with no code? Do you plan to release it?

开源期货？

It can not be reproduced. How can I know your work is true? You should make it open to public like Facebook and OpenAI. A demo at least.

阿里特色开源

This kinda start needs to start being removed from Github.

While the paper is very detailed, and will help anyone with the know how and compute to reproduce this, i think the idea of using SD1.5 as a starting base, and AD, kinda means you really should be releasing these models. You are building on the literal compute of others, and in the case, its most likely only fair to return the favor.

Combined with many of the concepts coming from the open source community, such as ReferenceNet with originated from the ControlNet devs.

Or at least dont make these sort of github where you just seeking exposure, this is a code repo after all.

The animation has lots of physical object interactions e.g. lighting, reflection in glass, movement of earrings etc.

These are highly unlikely with the current algorithmic training architecture stated in the paper.
Please release the code and preliminary model for us to verify.
Even boximator in TikTok cannot reproduce such quality. Both do not release the codes as usual and hide behind marketing.
It just show that such deepfakes are now high quality and users should be wary of adversarial attacks like this from now on.
Please feel free to contact us at Amaris.AI if you need defences against such deepfakes!

论文不是还在arxiv上吗？录用了再发代码也正常啊，怎么都开喷了捏

这个项目是想让从业者都EMO是吧，希望有真正开源的一天

虽然0代码开源，但我愿意再详细阿里一次，别再让我们失望

666666666666666666666666666666666666

怎么玩

这个怎么玩

RealEMO is here!!!

0代码还能2Kstar, 还能再假一点吗

至少放个infer的代码吧，然后模型给个效果差点的也行

don't understand why they upload it to GitHub without even opening the source

I don't understand why they choose not to make it public if it was created using open source. If that's the case, they should just keep it within Alibaba. Such selfish people

别开源好吗？

挂个github url，只开源个readme，求求别开源了好吗

垃圾，删了吧，恶心

恶心恶心恶心

好奇作者是不是用的Image-driven类+人脸高清的开源算法跑的，可不可以发个TTS驱动的效果

NO CODE!!!!!!

不敢开源代码模型的一律视为造假

不开还在这里开个github赚star，又当又立

仔细想想EMO的readme里并没有写它要开源代码啊

不像AnimateAnyone写了要开源但是并不打算开源

骗赞吗？

功能和这个差不多：用AI工具生成我奶奶的虚拟数字人

https://b23.tv/DrhGJJ9

大佬们，想开源就正经搞开源，别只放个演示视频

不要又当又立，说开源就整个空项目。man what can I say？

Don't upload to github without any code release! shame on you! ❌❌❌

another Animate Anyone 🫣🫣🫣 ?

code release time

thanks for your fabulous work , its really amazing and I couldn't wait to release code. could you please tell me when you're going to release the code 💥

建议阿里把这群骗子裁掉，有损公司名誉，丢**人的脸

rt

支持开源，

看来AnimateAnyone这套框架和花了大量时间整理沉淀下来的数据确实牛逼，努力就会有收获

就这？至少给个huggingface的下载链接撒

就这？至少给个huggingface的下载链接撒

Unreasonably high quality animations

This is for sure one of those researches that I'm sceptic about. The animation is just TO good to be img2video from audio. Muscles are moving, backgrounds are visible in the animation that is not visible in the image, 3D head rotations, while small, are impossibly good, hair jiggle corresponding to the laws of physics...

Nah... I believe it when I can try it and verify it. Until then it's to good to be true!

Fake

devs didn't make anything, this is fake.

阿里近一年，总爱搞这种论文级开源，发论文你发就完了，瞎吹开源刷KPI 保职位吗？

TALK IS CHEAP!!!

希望大佬可以根据论文复现，开源出来，阿里估计是不会开源了

。。。

可以开放一个gradio吗，学学国外的开源精神，哪怕是付费使用呢，别总是在实验室搞，国家需要你们

混github好多年了，第一次看到issues中几乎100%骂的。如果太没有效果，也不会骂；骂是因为觉得好玩有点效果但不开源。 阿里也确实是找骂，因为github本来是放代码的，你占个坑就是找骂。 最好是删除，弄一个project页面就好了。 本来整一个也不是特别难，也公开了算法，大家别骂了。

要论文，不够；
要代码，还不够；
要权重，还不够；
要数据，还不够；
要debug，还不够；
要tunning，还不够；
要......

哈哈哈哈哈
人家算法结构图给了，描述的也比较清楚，大家就不要骂了。差不多得了。

Code Release

Wow this is fantastic great work!

Are you planning on releasing the code?

直接删掉吧，占空间

Could you show we how to try out this MODEL？

Can't wait to try out this model, hoping to provide a locally testable model and corresponding methods ASAP.

Where can I download it?

Where can I download it?