Repository navigation
IC-Light V2 (Flux-based IC-Light models) #98
Replies: 73 comments 66 replies
|
Super! Thanks for your work! |
|
still not working well on full body photo...... |
|
In case people ask: You can only use the hf space to test for now. More inference codes and weights may be released later. However, unlike V1, the license will be non-commercial only.
|
|
https://discord.com/channels/1282119084145643520/1282119084749619334/1300276485449322627 |
Regarding license... will fal have a commercial license for their API? That could be a solution for many who want to use it commercially |
|
Oh so Fal is offering this and we get it from them? How would we use that in ComfyUI? |
|
Cant wait to get FAL access! |
Now you can: #98 (comment) :-) |
|
Thanks you! It was only 2 days ago I was thinking I really hope we get IC-Light for Flux. You are a legend :) |
|
I hope we can get sd3.5 soon |
|
Awesome work! Any tips on how to handle cases where some color preservation is needed? For example here the colors of the graphic print on the tank-top change too much: |
|
I think maybe you can segment a mask for the source content you want to be preserved (e.g. use SAM to segment the graphic print on the tank-top). Then when in the denoising steps of generation, you can substitute predicted latents with source latents within masked area. For more fine-grained control, you can use a weight hyperparameter to control the ratio of mixture between predicted and source latents. |
|
Very powerful! |
|
Thanks for the awesome work!
|
|
So awesome! It will be great when it can be applied to ComfyUI! |
|
+1 |
|
+1 |
|
Now you can: #98 (comment) :-) |
|
+1 |
|
Great job! Hope sometime I can use API to direct use it. |
|
is it applicable for comfyui rn? |
|
great work! on which flux model is it going to base on? schnell would be amazing in regards to its commercial license :) |
|
Awesome! |
|
I found that in the original picture, for example, a bottle of shampoo, the Chinese characters on the bottle become distorted |
|
解决了这个问题吗，同志是否有图片案例可分享，我想知道你的解决了这个问题是到了什么程度 |
|
试试4k以上分辨率，但是这样对硬件要求就会高了，时间也会长 |
|
后处理调用一下restore detail试试 |
|
Where's restore detail ? |
|
|
Amazing. Bit of a bummer on it being non-commercial tho, but great work nonetheless. |
|
awesome work can you also add that light direction circle like comfyui had to the gradio demo before releasing local version of the app? thank you so much
|
|
Or just add a way to provide a custom the light map |
|
very look forward to it as well! |
|
I want a prison |
|
Tell me, will this be added to Forge? |
|
no |
|
|
|
Please can I get a free website or application for cloth removal please🙏.
I tried some several apps but it didn't work out. I need a free and legit
cloth removal website and free applications. Thanks. It's for Educational
purpose nothing attached.
…On Sat, Apr 5, 2025, 12:07 AM ussainn ***@***.***> wrote:
7640010eedef1aba7b249d107aaeb65e.jpg (view on web)
<https://github.com/user-attachments/assets/ebb360f0-2e02-4b26-9583-b7bfeb110912>
—
Reply to this email directly, view it on GitHub
<#98 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/BRATWIGE7NLJS77XEC7AUIT2X4GCRAVCNFSM6AAAAABQVXRJS6VHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTENZTGEYDONQ>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
|
|
Superb |
|
Superb |
|
FYI: I just published a custom node for using V2 via fal.ai API in ComfyUI workflows. |
|
How can i get this app on my phone please |
|
They're private
…On Sun, Apr 27, 2025, 5:29 AM wanga022 ***@***.***> wrote:
How can i get this app on my phone please
—
Reply to this email directly, view it on GitHub
<#98 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/BRATWIEMW5WRKUHQ4IW3RDL23RMKZAVCNFSM6AAAAABQVXRJS6VHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTEOJVHA2TANQ>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
|
|
Is there a V2 model? Or is it just an SD1.5 pipeline? A marketing ploy. The model is positioned as open source, but is only available through the API for money) It seems like it's time to train your own model on your dataset and make it publicly available. |
|
What are the possibilities to use this model commercially? I am assuming that using the model via fal.ai doesn't mean you can use it commercially. |
|
什么时候开源？或者可以在哪里调用到可商业的API接口？ |
|
The hugginface space is gone (404) |
|
You r file |
|
FLUX.1 Kontext looks like it's a very promising alternative to this and there should be an open-weights release soon™. |
|
The downside is you can't use the flux-dev and the soon be released Kontext version commercially. |
|
I believe that goes for IC-Light V2 as well though (so not a downside in a direct comparison). On the other hand, Kontex dev will be open-weights and it can do a lot more than just change lighting. |
|
You said it would be released soon, but it is not here, and there is no guarantee that it will be released. |
|
@vivi-gomez It was just released a few hours ago: https://huggingface.co/black-forest-labs/FLUX.1-Kontext-dev |
|
Hello! Thank you for creating this great Space. The application does not start. The error in the logs is: NameError: name 'GRADIO_SERVER_NAME' is not defined. It looks like the code uses the variable GRADIO_SERVER_NAME, but it is not declared. |
|
Isn't working |
|
hi, still not working |
and others
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Note: IC-Light V2 HF Space is currently under maintaining
Note that this post is a work in progress (wip). Maybe I will edit it a lot recently.
IC-Light V2 is a series of Flux-based IC-Light models with 16ch VAE and native high resolution.
We plan to have several model variations:
News
See also alternative model for stronger illumination variations and modifications here: #109
Foreground Conditioned Model
https://huggingface.co/spaces/lllyasviel/iclight-v2
Some notes:
Thanks to 16ch VAE, native high resolution, and better training methods, the detail preservation capability is much higher than SD1.5.
For example:
If you overlay the inputs/outputs, you will see the level of detail preservation, like:
another one:
Another important difference to SD1.5 is the ability to process stylized images. The SD15 IC-Light sometimes makes too large modification to image styles, like:
As you can see, the SD15 iclight makes the input drawing a photo-like image. The effect is still interesting but may not be the desired one in some cases.
Below is another attempt to add “surrealist painting of” to prompt but still cannot get a painting:
These are from new model:
(3 results with same seed 12345)
It depends on the taste of different people.
But at least the current model is finally able to give you a painting when you do not want a photo..
Examples, all with seed 12345
More info about other variations to be continued ...
Special thanks
Special thanks to FAL for the help with computation resources on many internal experiments. (I will edit this "special thanks" section later to add more items...)
All reactions