Software Engineering Courses (SECourses)•10mo ago

<@205854764540362752> is it possible to fine tune lora or full checkpoint of the FLUX Fill model tha

@Furkan Gözükara SECourses is it possible to fine tune lora or full checkpoint of the FLUX Fill model that is used for inpainting, to inpaint a particular thing like a logo or such?

Mm3dia_offline <@205854764540362752> is it possible to fine tune lora or full checkpoint of the...

Furkan Gözükara SECourses•3/4/25, 10:37 AM

yes

Furkan Gözükara SECourses•3/4/25, 10:37 AM

https://github.com/nftblackmagic/catvton-flux

GitHub

GitHub - nftblackmagic/catvton-flux

Contribute to nftblackmagic/catvton-flux development by creating an account on GitHub.

Gill Bastar•3/4/25, 7:34 PM

Is it possible in Wan 2.1 to repeatedly generate the same video and change only small details using prompts? For example, let's say I have a scene and I only don't like some object in the background, which I want to prompt differently. Or the resolution. Everything else is perfect and should not change. Or is the output always random?

And is it possible to fine-tune Wan 2.1?

GGill Bastar Is it possible in Wan 2.1 to repeatedly generate the same video and change only ...

Furkan Gözükara SECourses•3/4/25, 8:02 PM

it supports multi line input

Furkan Gözükara SECourses•3/4/25, 8:02 PM

use same seed to test

Furkan Gözükara SECourses•3/4/25, 8:02 PM

also lora training possible

Furkan Gözükara SECourses•3/4/25, 8:02 PM

i am gonna make a tutorial about it hopefully

Dazzastrous [4090]•3/5/25, 7:28 AM

doc can u implement an option to save settings / parameters in wan ?

DDazzastrous [4090]doc can u implement an option to save settings / parameters in wan ?

Furkan Gözükara SECourses•3/5/25, 11:02 AM

today i plan to add it

Furkan Gözükara SECourses•3/5/25, 11:02 AM

you reminded me good

herocus810•3/5/25, 1:22 PM

Hi Doc I just follow your installation for python, cuda, cudnn etc... which version of cudnn should I install?

Hherocus810 Hi Doc I just follow your installation for python, cuda, cudnn etc... which vers...

Furkan Gözükara SECourses•3/5/25, 1:48 PM

8.9 and 9.7

Furkan Gözükara SECourses•3/5/25, 1:48 PM

have both

herocus810•3/5/25, 1:49 PM

ok thanks, the one is currently is 9.8 no good?

Hherocus810 ok thanks, the one is currently is 9.8 no good?

Furkan Gözükara SECourses•3/5/25, 1:51 PM

i think should be fine instead of 9.7

Furkan Gözükara SECourses•3/5/25, 1:51 PM

but i didnt use it yet

herocus810•3/5/25, 1:52 PM

ok thanks, the one need to instal is 8.9.7 for cuda 12.x ya?

Hherocus810 ok thanks, the one need to instal is 8.9.7 for cuda 12.x ya?

Furkan Gözükara SECourses•3/5/25, 1:59 PM

8.9 still may be required some older apps

Furkan Gözükara SECourses•3/5/25, 1:59 PM

but it is all moving to 9.4+

Furkan Gözükara SECourses•3/5/25, 2:00 PM

slowly

Schmit•3/5/25, 3:20 PM

does anyone know a tool that can replace magnific perfectly?

SSchmit does anyone know a tool that can replace magnific perfectly?

Furkan Gözükara SECourses•3/5/25, 6:04 PM

Supir is best to upscale with adding real details

Furkan Gözükara SECourses•3/5/25, 6:05 PM

We have it

Furkan Gözükara SECourses•3/5/25, 6:05 PM

Actually currently updating it for rtx 5000 series

Dazzastrous [4090]•3/5/25, 6:11 PM

https://huggingface.co/spaces/nupurkmr9/SynCD

SynCD - a Hugging Face Space by nupurkmr9

DDazzastrous [4090]https://huggingface.co/spaces/nupurkmr9/SynCD

Furkan Gözükara SECourses•3/5/25, 6:14 PM

Did you test

Furkan Gözükara SECourses•3/5/25, 6:14 PM

Examples are looking like cherry pick

Dazzastrous [4090]•3/5/25, 6:45 PM

yeh i tested a few images worked good

DDazzastrous [4090]yeh i tested a few images worked good

Furkan Gözükara SECourses•3/5/25, 7:14 PM

i tested myself and haaha

Furkan Gözükara SECourses•3/5/25, 7:15 PM

art-ai_usa-girl•3/5/25, 7:26 PM

LTX-Video v0.9.5 released, now with keyframes, video extension, and higher resolutions support. https://github.com/Lightricks/LTX-Video

GitHub

GitHub - Lightricks/LTX-Video: Official repository for LTX-Video

Official repository for LTX-Video. Contribute to Lightricks/LTX-Video development by creating an account on GitHub.

Aart-ai_usa-girl LTX-Video v0.9.5 released, now with keyframes, video extension, and higher resol...

Furkan Gözükara SECourses•3/5/25, 7:27 PM

but it looks weak

Furkan Gözükara SECourses•3/5/25, 7:27 PM

examples

Aikage•3/6/25, 2:19 AM

Sdxl 3.5 turbo out

Aikage•3/6/25, 2:19 AM

For the three people not using flux

Vee•3/6/25, 5:01 AM

hey, I wanted to know if anyone has thoughts on this. If i want to create a singular animated character that I can provide text to and it says it and provide expressions based on the text.

is this even a feasible project?

AAikage For the three people not using flux

Furkan Gözükara SECourses•3/6/25, 8:37 AM

i feel like it is worse than 3.5 large did you notice?

VVee hey, I wanted to know if anyone has thoughts on this. If i want to create a sing...

Furkan Gözükara SECourses•3/6/25, 8:37 AM

you have to train a model for that to work

herocus810•3/6/25, 11:31 AM

Hi Doc, when I run comfyui I've got an error message... any idea what caused this? But the comfyui keep running and open tproperly with no problem after I clik ok

herocus810•3/6/25, 11:31 AM

cyril23•3/6/25, 12:12 PM

Tencent Releases HunyuanVideo-I2V

see https://www.reddit.com/r/StableDiffusion/comments/1j4qrh8/comment/mgav9qr/?utm_source=share&utm_medium=mweb3x&utm_name=mweb3xcss&utm_term=1&utm_content=share_button

comfyui workflow: https://github.com/kijai/ComfyUI-HunyuanVideoWrapper/blob/main/example_workflows/hyvideo_i2v_example_01.json

download quant hunyuan_video_I2V_fp8_e4m3fn.safetensors from https://huggingface.co/Kijai/HunyuanVideo_comfy/tree/main
(see https://www.reddit.com/r/StableDiffusion/comments/1j4qrh8/comment/mgav9qr/?utm_source=share&utm_medium=web3x&utm_name=web3xcss&utm_term=1&utm_content=share_button)

General instructions for comfyui and hunyuan video (especially how to build SageAttention under windows): https://www.reddit.com/r/StableDiffusion/comments/1h7hunp/how_to_run_hunyuanvideo_on_a_single_24gb_vram_card/

Ccyril23 Tencent Releases HunyuanVideo-I2V see https://www.reddit.com/r/StableDiffusion/...

Dazzastrous [4090]•3/6/25, 1:47 PM

Another sleepless night for Doc hehe.......

Hherocus810 Click to see attachment

Furkan Gözükara SECourses•3/6/25, 3:17 PM

i also seen this error

Furkan Gözükara SECourses•3/6/25, 3:17 PM

i dont know exact reason yet

FFurkan Gözükara SECourses i also seen this error

herocus810•3/6/25, 3:18 PM

yes it's weird

A. T.•3/6/25, 3:56 PM

Hi. I got these errors while installing Windows_Install_ALL_GPU for WAN21_v26

A. T.•3/6/25, 3:56 PM

message.txt3.23KB

A. T.•3/6/25, 3:56 PM

A. T.•3/6/25, 4:23 PM

and

C:\AI\WAN_2.1>Start_APP.bat
The system cannot find the path specified.

C:\AI\WAN_2.1>Start_APP.bat
The system cannot find the path specified.

relative to :

call .\venv\Scripts\activate.bat || exit /bcall .\venv\Scripts\activate.bat || exit /b

AA. T.Hi. I got these errors while installing Windows_Install_ALL_GPU for WAN21_v26

Furkan Gözükara SECourses•3/6/25, 6:00 PM

hi let me update and fix

<@205854764540362752> is it possible to fine tune lora or full checkpoint of the FLUX Fill model tha

Similar Threads

Similar Threads

Similar Threads