Software Engineering Courses (SECourses)•2y ago

[Tiled VAE]: Executing Encoder Task Queue: 100%|████████████████████████████████████████████████████

[Tiled VAE]: Executing Encoder Task Queue: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 5733/5733 [00:19<00:00, 287.36it/s]
[Tiled VAE]: Done in 20.535s, max VRAM alloc 10024.335 MB

says its done but not got any output just the orange shapes

Rrafraf `The image features a close-up of a woman with short brown hair and blue eyes, w...

leem0nchu•3/1/24, 1:38 AM

A tip that I don’t know if it will apply in your case. But I do most of my captions with Kosmos. I found that if you give a brief description (I do it for batches) eg; “This is an image of a [giant]. Describe the image in detail….

It will eliminate things like The image features, depicts, In the image, etc, etc.

DDazzastrous [4090][Tiled VAE]: Executing Encoder Task Queue: 100%|████████████████████████████████...

Furkan Gözükara SECourses•3/1/24, 1:39 AM

weird

Furkan Gözükara SECourses•3/1/24, 1:39 AM

check outputs folder

Furkan Gözükara SECourses•3/1/24, 1:39 AM

if saved

Lleem0nchu A tip that I don’t know if it will apply in your case. But I do most of my capti...

Furkan Gözükara SECourses•3/1/24, 1:39 AM

nice tips

FFurkan Gözükara SECourses if saved

Dazzastrous [4090]OP•3/1/24, 1:40 AM

Its popped up at last

Lleem0nchu A tip that I don’t know if it will apply in your case. But I do most of my capti...

rafraf•3/1/24, 1:42 AM

but does it matter that much? Did you see a difference when you dont have "The image features"?

Dazzastrous [4090]OP•3/1/24, 1:42 AM

didnt save into the outputs folder though

Rrafraf but does it matter that much? Did you see a difference when you dont have "The i...

leem0nchu•3/1/24, 1:43 AM

For training it does as captioning is so important as image quality.

DDazzastrous [4090]didnt save into the outputs folder though

Furkan Gözükara SECourses•3/1/24, 1:44 AM

it should save

Furkan Gözükara SECourses•3/1/24, 1:44 AM

did you define any other outputs?

Dazzastrous [4090]OP•3/1/24, 1:44 AM

nope

Lleem0nchu For training it does as captioning is so important as image quality.

rafraf•3/1/24, 1:44 AM

it's for supir upscaling prompt

Furkan Gözükara SECourses•3/1/24, 1:44 AM

check outputs again

Dazzastrous [4090]OP•3/1/24, 1:44 AM

lol delayed its now there

leem0nchu•3/1/24, 1:44 AM

Ah well I that case I couldn’t tell I haven’t used it for that purpose.

Dazzastrous [4090]OP•3/1/24, 1:47 AM

nice

Dazzastrous [4090]OP•3/1/24, 1:48 AM

heheh its his arm!!!

Dazzastrous [4090]OP•3/1/24, 1:55 AM

Dr have you got the link to the image slider?

Furkan Gözükara SECourses•3/1/24, 2:16 AM

the links are on the patreon post

VAMP•3/1/24, 2:16 AM

@DoctorCoach you need to check out EMO. . It's the heygen alternative that we have been looking for ..see the demo videos..it's crazy..it captures all the expressions , micro expressions, mouth,lip movement, eye movement....
https://youtu.be/4znFrPcnRnY

YouTubeMark Liberman

AlibabaSong1

From Linrui Tian, Qi Wang, Bang Zhang, & Liefeng Bo, "EMO: Emote Portrait Alive - Generating Expressive Portrait Videos with Audio2Video Diffusion Model under Weak Conditions", https://humanaigc.github.io/emote-portrait-alive/.

VAMP•3/1/24, 2:16 AM

just waiting for its release

VVAMP @DoctorCoach you need to check out EMO. . It's the heygen alternative that we h...

Furkan Gözükara SECourses•3/1/24, 2:18 AM

probably theyt wont release ever

Furkan Gözükara SECourses•3/1/24, 2:18 AM

ali baba does that

VAMP•3/1/24, 2:18 AM

They are listed on GitHub

VVAMP They are listed on GitHub

Furkan Gözükara SECourses•3/1/24, 2:18 AM

they did this before

VAMP•3/1/24, 2:18 AM

Hmm

Dazzastrous [4090]OP•3/1/24, 2:19 AM

Ah yeh Ali baba i recognise that name

Furkan Gözükara SECourses•3/1/24, 2:19 AM

https://github.com/HumanAIGC/AnimateAnyone

GitHub

GitHub - HumanAIGC/AnimateAnyone: Animate Anyone: Consistent and Co...

Animate Anyone: Consistent and Controllable Image-to-Video Synthesis for Character Animation - HumanAIGC/AnimateAnyone

Dazzastrous [4090]OP•3/1/24, 2:19 AM

yup thats the one

Furkan Gözükara SECourses•3/1/24, 2:19 AM

13k stars 0 code

Furkan Gözükara SECourses•3/1/24, 2:19 AM

and alibaba is 100s billions dolar company

Dazzastrous [4090]OP•3/1/24, 2:19 AM

yeh i looked into that

Dazzastrous [4090]OP•3/1/24, 2:19 AM

very odd

FFurkan Gözükara SECourses 13k stars 0 code

VAMP•3/1/24, 2:19 AM

Then what's the point of listing these things on GitHub .? When they never release their software or code

VVAMP Then what's the point of listing these things on GitHub .? When they never relea...

Furkan Gözükara SECourses•3/1/24, 2:20 AM

show off just like google demos

Dazzastrous [4090]OP•3/1/24, 2:20 AM

gets the name out there

VAMP•3/1/24, 2:20 AM

What are they trying to do? Show smart they are

Furkan Gözükara SECourses•3/1/24, 2:20 AM

FFurkan Gözükara SECourses show off just like google demos

VAMP•3/1/24, 2:20 AM

Okay

FFurkan Gözükara SECourses show off just like google demos

VAMP•3/1/24, 2:23 AM

@Dr. Furkan Gözükara
https://youtu.be/ba65xD_0DfQ
Demos are really good..Like better than all other deepfakes...

YouTubeMeta Professor

EMO-Emote Portrait Alive - Generate Video through AI (Audio+Image -...

Generate a talking head from a single photo, with emotion and movement.

Alibaba group just dropped EMO (Emote Portrait Alive). All you need is a single image and an audio file (both of which can be generated using other AI frameworks).

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