Software Engineering Courses (SECourses)•16mo ago

thats gonna take a while 🙂

thats gonna take a while

Ddaroth thats gonna take a while 🙂

Furkan Gözükara SECourses•9/21/24, 6:16 PM

Maxivi•9/21/24, 6:20 PM

@Dr. Furkan Gözükara any update on the fp8 for dreambooth Furak Obama?

Maxivi•9/21/24, 6:31 PM

@Dr. Furkan Gözükara do you get this issue when extracting lora

flux_extract_lora.py:68
31%|█████████████████████████████████████████████████████████████████████████████ | 93/304 [01:57<09:26, 1.86s/it]
Traceback (most recent call last):
File "/workspace/kohya_ss/sd-scripts/networks/flux_extract_lora.py", line 219, in <module>
svd(**vars(args))
File "/workspace/kohya_ss/sd-scripts/networks/flux_extract_lora.py", line 87, in svd
value_t = f_tuned.get_tensor(key)
File "/workspace/kohya_ss/sd-scripts/library/utils.py", line 249, in get_tensor
return self._deserialize_tensor(tensor_bytes, metadata)
File "/workspace/kohya_ss/sd-scripts/library/utils.py", line 265, in _deserialize_tensor
byte_tensor = torch.frombuffer(tensor_bytes, dtype=torch.uint8)
ValueError: both buffer length (0) and count (-1) must not be 0

MMaxivi @Dr. Furkan Gözükara do you get this issue when extracting lora ...

Furkan Gözükara SECourses•9/21/24, 6:36 PM

nope

MMaxivi @Dr. Furkan Gözükara any update on the fp8 for dreambooth Furak Obama?

Furkan Gözükara SECourses•9/21/24, 6:36 PM

nope too

Furkan Gözükara SECourses•9/21/24, 6:36 PM

testing first checkpoint

Furkan Gözükara SECourses•9/21/24, 10:22 PM

looks like model collapsed gonna test 2 different trainings more and also will wait checkpoint 2

Furkan Gözükara SECourses•9/21/24, 11:40 PM

new thing arrived

Furkan Gözükara SECourses•9/21/24, 11:40 PM

https://www.reddit.com/r/StableDiffusion/comments/1fm9pxa/joycaption_free_open_uncensored_vlm_alpha_one/

From the StableDiffusion community on Reddit: JoyCaption: Free, Ope...

Explore this post and more from the StableDiffusion community

Furkan Gözükara SECourses•9/21/24, 11:40 PM

i will try to add all new features to my captioner

Furkan Gözükara SECourses•9/22/24, 12:05 AM

i think rtx a6000 can do batch size 8

Furkan Gözükara SECourses•9/22/24, 12:05 AM

when doing fine tuning .d

Furkan Gözükara SECourses•9/22/24, 12:05 AM

verifyinmg it right now

Furkan Gözükara SECourses•9/22/24, 12:08 AM

yep doing 8x amazing

Furkan Gözükara SECourses•9/22/24, 12:09 AM

dmn step 2 failed :d

Furkan Gözükara SECourses•9/22/24, 12:09 AM

i guess max is 7

Furkan Gözükara SECourses•9/22/24, 12:11 AM

testng with cpu offloading

xlads•9/22/24, 12:27 AM

xlads•9/22/24, 12:28 AM

I genuinely think flux is getting better

xlads•9/22/24, 12:28 AM

I didn’t even differentiate when there was multiple people in the scene, but it made the people look different

Furkan Gözükara SECourses•9/22/24, 1:15 AM

alpha1

xlads•9/22/24, 1:35 AM

@Dr. Furkan Gözükara you are a literal goat

xlads•9/22/24, 1:35 AM

The e.t’s sent you here for a reason

Xxlads Click to see attachment

Marrvelous•9/22/24, 1:35 AM

What did you use to do the video?

xlads•9/22/24, 1:36 AM

Runwayml new update!

xlads•9/22/24, 1:38 AM

Vertical option is available

Xxlads @Dr. Furkan Gözükara you are a literal goat

Furkan Gözükara SECourses•9/22/24, 1:38 AM

thanks

xlads•9/22/24, 1:38 AM

And the new open AIGPT01

Furkan Gözükara SECourses•9/22/24, 1:38 AM

i am trying to add 4bit quant right now

Furkan Gözükara SECourses•9/22/24, 1:58 AM

wow 4 bit saves 10 gb

Furkan Gözükara SECourses•9/22/24, 2:11 AM

4 bit uses 8500 mb only

xlads•9/22/24, 3:06 AM

Live portrait for video @Dr. Furkan Gözükara can you help out here?

Xxlads Live portrait for video @Dr. Furkan Gözükara can you help out here?

Furkan Gözükara SECourses•9/22/24, 3:14 AM

what help

Furkan Gözükara SECourses•9/22/24, 3:19 AM

doing a lot of experiments for big dataset training

Furkan Gözükara SECourses•9/22/24, 3:19 AM

Furkan Gözükara SECourses•9/22/24, 3:19 AM

10400 , 20800

Furkan Gözükara SECourses•9/22/24, 3:20 AM

joy caption updated : 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•9/22/24, 3:23 AM

i am going to sleep hopefully see you tomorrow

Maxivi•9/22/24, 10:02 AM

@Dr. Furkan Gözükara on your new joy caption app, the caption length does not work. for example if you set it to 70 its uses more

Maxivi•9/22/24, 10:03 AM

For RNG tags it works fine
For training_prompt it does not work
For descriptive it does not work

MMaxivi @Dr. Furkan Gözükara on your new joy caption app, the caption length does not wo...

Furkan Gözükara SECourses•9/22/24, 12:13 PM

well it is authors error not mine :d

Furkan Gözükara SECourses•9/22/24, 12:13 PM

i am adding custom prompt option as well

Maxivi•9/22/24, 12:13 PM