Software Engineering Courses (SECourses)•3y ago

Hello, Everyone. I'm having an issue. If someone knows, could they please help me? Thanks! I don't

Hello, Everyone. I'm having an issue. If someone knows, could they please help me? Thanks!

I don't exactly know what is causing this. There is no error per se. I'm trying to use the SOTA Image Captioning Scripts and Clip_Interrogator Gradio Web UI but it seems much slower than the demonstration. I also have a 3090 so I should be getting the same speeds.

I'm using
ViT-bigG-14/laion2b_s39b_b160k
blip2-2.7b

Though, this is the current ETA:

SarcasticestOP•12/13/23, 2:17 AM

EyeSpyBekoAI•12/13/23, 2:18 AM

Hello everyone!
I had a quick question for you. I seem to remember coming across conversations here with respect to having stable diffusion model(s) that allowed stylized text (word) generation, without the use of control net. For the life of me though I can't remember what or where it was. Do you know if there is such a thing? Thanks for your help!

EEyeSpyBekoAI Hello everyone! I had a quick question for you. I seem to remember coming across...

Lorelai AI•12/13/23, 2:20 AM

https://civitai.com/models/176555/harrlogos-xl-finally-custom-text-generation-in-sd

Harrlogos XL - Finally, custom text generation in SD! - Harrlogos_v...

Introducing Harrlogos v2.0! Solving Stable Diffusion text generation one LoRA at a time! THIS DOES NOT REQUIRE CONTROL NET TO WORK Harrlogos now on...

Lorelai AI•12/13/23, 2:20 AM

There's two out now I think, you need to use SDXL tho

LLorelai AI There's two out now I think, you need to use SDXL tho

EyeSpyBekoAI•12/13/23, 2:29 AM

Thank you!!!
No worries. I’m fully on SDXL

EEyeSpyBekoAI Thank you!!! No worries. I’m fully on SDXL

Lorelai AI•12/13/23, 2:30 AM

I hope everyone gets to

LLorelai AI I hope everyone gets to 🙂

EyeSpyBekoAI•12/13/23, 2:32 AM

With the exception of doing sdxl training, is there something that stops you from using sdxl?
Maybe I can help?

EEyeSpyBekoAI With the exception of doing sdxl training, is there something that stops you fro...

Lorelai AI•12/13/23, 2:33 AM

I only use SDXL, all day literally

LLorelai AI I only use SDXL, all day literally

EyeSpyBekoAI•12/13/23, 2:34 AM

Got ya. I interpreted your message a different way I think so thought I could help. Thanks again!

EEyeSpyBekoAI Got ya. I interpreted your message a different way I think so thought I could he...

Lorelai AI•12/13/23, 2:34 AM

No problem nice to meet you!

Lorelai AI•12/13/23, 2:35 AM

Harrlogos isnt perfect but its pretty cool

Lorelai AI•12/13/23, 2:35 AM

Lorelai AI•12/13/23, 2:36 AM

Lorelai AI•12/13/23, 2:36 AM

The bottom one I modified

LLorelai AI No problem nice to meet you!

EyeSpyBekoAI•12/13/23, 2:49 AM

Nice to meet you too.
These are great! Looks like I’ll be up late tonight

EEyeSpyBekoAI Nice to meet you too. These are great! Looks like I’ll be up late tonight 🙂

Lorelai AI•12/13/23, 2:50 AM

me too probably, I just did a merge and I'm going to build a tensor engine for it, I pray that the SDXL Style selector works then I'll have what I want. The merge went great.

LLorelai AI me too probably, I just did a merge and I'm going to build a tensor engine for i...

EyeSpyBekoAI•12/13/23, 3:02 AM

Oh cool you build as well!
I don’t build tensor engines but I am working on projects merging several AI concepts. Happy to DM and chat more if you’d like to chat more.

EEyeSpyBekoAI Oh cool you build as well! I don’t build tensor engines but I am working on proj...

Lorelai AI•12/13/23, 4:02 AM

We can chat here it's all good

andysingal•12/13/23, 6:40 AM

Christmas vibes using Stable diffusion: https://youtu.be/iqv2Xn0UVnA?si=BRbw85Z5U6PNZl0a

YouTubeAndyAI Insights

A Festive Journey: Embracing the Spirit of Christmas #ai #technolog...

Dive into the enchanting world of Christmas with our heartwarming video that captures the essence of the holiday season. Join us on a magical journey filled with twinkling lights, joyous melodies, and the spirit of giving.

As snowflakes gently fall, families gather around crackling fires, sipping hot cocoa and sharing laughter. The streets are ...

BBophedes i have same issue, where did you put this command?

Furkan Gözükara SECourses•12/13/23, 9:36 AM

hello did you fix your error?

SSarcasticest Hello, Everyone. I'm having an issue. If someone knows, could they please help m...

Furkan Gözükara SECourses•12/13/23, 9:37 AM

yes you are getting into shared vram

Furkan Gözükara SECourses•12/13/23, 9:37 AM

something must be wrong

Furkan Gözükara SECourses•12/13/23, 9:37 AM

something else on your computer must be using your VRAM

LLorelai AI https://civitai.com/models/176555/harrlogos-xl-finally-custom-text-generation-in...

Furkan Gözükara SECourses•12/13/23, 9:38 AM

dmn this is nice lora

LLorelai AI Click to see attachment

Furkan Gözükara SECourses•12/13/23, 9:39 AM

this is great

FFurkan Gözükara SECourses this is great

Lorelai AI•12/13/23, 9:40 AM

I used Adobe to fix the R and add the fun dice. I supplement my pictures there to put the finishing touches

LLorelai AI https://civitai.com/models/176555/harrlogos-xl-finally-custom-text-generation-in...

EyeSpyBekoAI•12/13/23, 9:43 AM

@The MisoGenius , it's almost 2am here and I've been playing with this. I will say, it is nice, but it does have limitations. If a word is more than 4-5 letters, it starts getting confused. If it's a popular phrase though like "happy new year", it'll do better. Definitely not an "end all be all" model, but it does get you started. Thanks again for pointing me to it. Here is something I generated. It did take a lot of inpainting though as it can't follow this many letters, especially if you decide to change up the spelling. But, overall a really nice way to get you going and creative.

EEyeSpyBekoAI @The MisoGenius , it's almost 2am here and I've been playing with this. I will s...

Lorelai AI•12/13/23, 9:46 AM

What I will say is the technology will become capable of anything but the art is limitless

EyeSpyBekoAI•12/13/23, 9:46 AM

Yea absolutely the art is stellar I have nothing to say towards that. It's incredibly creative.

EEyeSpyBekoAI Yea absolutely the art is stellar I have nothing to say towards that. It's incre...

Lorelai AI•12/13/23, 9:49 AM

I'm launching my site probably tomorrow along with my socials im just wrapping up its why im up late. The project I felt like doing earlier was I trained my subject but i captioned him as a woman and I used women for Class images because I wanted to use loras for womens clothing and to make it more versatile. And I was right.

LLorelai AI I'm launching my site probably tomorrow along with my socials im just wrapping u...

EyeSpyBekoAI•12/13/23, 9:49 AM

Similar. I'm hoping to get my site out this week.

EEyeSpyBekoAI Similar. I'm hoping to get my site out this week.

Lorelai AI•12/13/23, 9:50 AM

Sweeettttt

FFurkan Gözükara SECourses 3.10.11 is best i think

nissanle•12/13/23, 9:56 AM

Thank you Furkan for your help.

nannigath•12/13/23, 10:23 AM

decorify.app_mid-century_living_room.png

nannigath•12/13/23, 10:25 AM

I recently discovered Decorify (https://decorify.app/) and tested it with a living room image. Unlike my attempts with ControlNet, which maintained the furniture's structure, Decorify generated images with varied styles while preserving the room shape and rearranging the furniture in different shapes, like the diverse coffee table designs in each image. What might be their approach, and how does it differ from ControlNet?

nannigath•12/13/23, 10:26 AM

So far I have tried controlNet with mlsd, depth , segmentation and canny

JeremytheJedi•12/13/23, 10:32 AM

So i had a chance to go through the Zero to Hero Stable Diffusion YouTube (great video btw). Althought our versions are slightly different, I was able to identify and set all the settings the way you did. Unfortunately, when I go to train it, it's just generating black pictures. So... I might be missing something. Any thoughts?