Software Engineering Courses (SECourses)•17mo ago

Hey @Dr. Furkan Gözükara 🙂 I’m trying to install CUDA 11.8 on RTX 4090 for training flux. Every ti

Hey @Dr. Furkan Gözükara

I’m trying to install CUDA 11.8 on RTX 4090 for training flux.
Every time I try to install it fails. I’ve tried a couple of different versions but still no luck.
I’m using windows 11 pro, with Nvidia studio driver version 531.41

Nnicearoni i've done a flux lora with "ohwx man" and another with captions generated by Joy...

amyn•8/25/24, 1:04 PM

I myself have glasses and trained a lora with 20/25 images with glasses, it always puts them by default. Have not used captions tho because I trained using ai-toolkit on replicate.

One issue I have isit always puts my face on other people in the background too and is heavily influenced by the background I had on the input training images (old city buildings)

KGTachi•8/25/24, 1:05 PM

@Dr. Furkan Gözükara can you please add the option in joy caption to cut off the last sentence. It’s always cut off at the end. So and option to cut of the last one would be perfect.

Aamyn I myself have glasses and trained a lora with 20/25 images with glasses, it alwa...

nicearoni•8/25/24, 1:08 PM

interesting. maybe a bit overtrained then? you could try a version of the lora based off fewer epochs\steps if ai-toolkit saves them along the way? or specify a background in your prompt?

Nnicearoni interesting. maybe a bit overtrained then? you could try a version of the lora b...

amyn•8/25/24, 1:20 PM

Yes totally, when specifying backroung in prompt it does work as intended

amyn•8/25/24, 1:20 PM

I used 2200 steps in training v2, and 1000 for v1

amyn•8/25/24, 1:20 PM

Both are the same in results tbh

smoke•8/25/24, 1:28 PM

Does anyone have different/worse results with Flux LoRA’s in Comfy than in Forge for example?

Ssmoke Does anyone have different/worse results with Flux LoRA’s in Comfy than in Forge...

�

𝔸𝖨ΞӾ•8/25/24, 3:15 PM

yeah when I tested loras out in forge they wouldn't have any effect at all most of the time, even with the trigger word in prompt, but in SwarmUI they worked flawlessly so I dunno what's happening there

TOPUI•8/25/24, 3:15 PM

After the installation is complete, press 1 to start the error

TOPUI•8/25/24, 3:15 PM

TOPUI•8/25/24, 3:16 PM

@Dr. Furkan Gözükara

PPoppy Hey @Dr. Furkan Gözükara 🙂 I’m trying to install CUDA 11.8 on RTX 4090 for trai...

Furkan Gözükara SECourses•8/25/24, 4:18 PM

how it fails exactly? nvidia has setup

KKGTachi @Dr. Furkan Gözükara can you please add the option in joy caption to cut off the...

Furkan Gözükara SECourses•8/25/24, 4:18 PM

sure let me try that option

Furkan Gözükara SECourses•8/25/24, 4:20 PM

i will also add multi gpu to joy caption app

TTOPUI Click to see attachment

Furkan Gözükara SECourses•8/25/24, 4:21 PM

i need to see install logs

Furkan Gözükara SECourses•8/25/24, 4:21 PM

reinstall and send me logs

KGTachi•8/25/24, 6:47 PM

@Dr. Furkan Gözükara would you also be making a internvl2 gradio app? Seems that’s the best caption model right now

Kallamamran•8/25/24, 8:15 PM

Tried some comparisons between samplers and schedulers on my new LORA

Kallamamran•8/25/24, 8:25 PM

All other combinations except these were somewhat useless!

KKGTachi @Dr. Furkan Gözükara can you please add the option in joy caption to cut off the...

Furkan Gözükara SECourses•8/25/24, 8:36 PM

feature added with v10 : https://www.patreon.com/posts/110613301

Patreon

JoyCaption : Amazing Image Captioner - Batch Processing Included - ...

Get more from SECourses: Tutorials, Guides, Resources, Training, FLUX, MidJourney, Voice Clone, TTS, ChatGPT, GPT, LLM, Scripts by Furkan Gözü on Patreon

Furkan Gözükara SECourses•8/25/24, 8:36 PM

KKGTachi @Dr. Furkan Gözükara would you also be making a internvl2 gradio app? Seems that...

Furkan Gözükara SECourses•8/25/24, 8:36 PM

any demo of this like on hugging face?

KKallamamran Tried some comparisons between samplers and schedulers on my new LORA

Furkan Gözükara SECourses•8/25/24, 8:36 PM

i think unipc normal best

Kallamamran•8/25/24, 8:37 PM

Yes it's good

Kallamamran•8/25/24, 8:38 PM

Normally... Doesn't work at all now

dsienra•8/25/24, 8:38 PM

I posted on github asking for Flux CLIP_L Text Encoder

Kallamamran•8/25/24, 8:38 PM

Don't know why

Kallamamran•8/25/24, 8:39 PM

Heun+Simple was really good now

dsienra•8/25/24, 8:39 PM

I was asking about CLIP_L but hes talking about T5... adk

Ddsienra I was asking about CLIP_L but hes talking about T5... adk

Maxivi•8/25/24, 8:41 PM

already spoke to sd scripts kohya, he is is working on clip l text encoder training

Maxivi•8/25/24, 8:41 PM

t5 will be frozen at all times

FFurkan Gözükara SECourses any demo of this like on hugging face?

Maxivi•8/25/24, 8:42 PM

think they're talking about this one mentioned earlier - https://huggingface.co/spaces/OpenGVLab/InternVL

InternVL - a Hugging Face Space by OpenGVLab

MMaxivi think they're talking about this one mentioned earlier - https://huggingface.co/...

Furkan Gözükara SECourses•8/25/24, 8:43 PM

it is api based

Furkan Gözükara SECourses•8/25/24, 8:43 PM

there is no code at that app

MMaxivi t5 will be frozen at all times

dsienra•8/25/24, 8:43 PM

yes T5 is too large

Furkan Gözükara SECourses•8/25/24, 8:45 PM

i am adding multi gpu to joycaption app

MMaxivi already spoke to sd scripts kohya, he is is working on clip l text encoder train...

Furkan Gözükara SECourses•8/25/24, 8:45 PM

yep

Ddsienra yes T5 is too large

Furkan Gözükara SECourses•8/25/24, 8:45 PM

even flux authors didnt train it

Furkan Gözükara SECourses•8/25/24, 8:45 PM

so unnecessary

FFurkan Gözükara SECourses it is api based

KGTachi•8/25/24, 8:53 PM

https://huggingface.co/OpenGVLab/InternVL2-8B/tree/main

OpenGVLab/InternVL2-8B at main

KGTachi•8/25/24, 8:53 PM

This one here

FFurkan Gözükara SECourses i am adding multi gpu to joycaption app

Kallamamran•8/25/24, 8:55 PM

Joycaption... Gotta try that one

Kallamamran•8/25/24, 8:55 PM

Is it fast?

KKallamamran Is it fast?

Furkan Gözükara SECourses•8/25/24, 8:59 PM

decent under 10 seconds

Furkan Gözükara SECourses•8/25/24, 8:59 PM

i will caption dataset on 8 gpu

Furkan Gözükara SECourses•8/25/24, 8:59 PM

trying to code multi gpu

Kallamamran•8/25/24, 9:00 PM

He he... I only have one

KKGTachi https://huggingface.co/OpenGVLab/InternVL2-8B/tree/main

Furkan Gözükara SECourses•8/25/24, 9:00 PM

it doesnt have example gradio there

hunterwolf602•8/25/24, 9:06 PM

Good evening! When I create a pod through RunPod using the Stable Diffusion template with web UI, But when connecting the pod, the option to access Jupyter Lab is not ready.

Hhunterwolf602 Good evening! When I create a pod through RunPod using the Stable Diffusion temp...

Furkan Gözükara SECourses•8/25/24, 9:11 PM

refresh and try again

Hey @Dr. Furkan Gözükara 🙂 I’m trying to install CUDA 11.8 on RTX 4090 for training flux. Every ti

Similar Threads