Replies: 23 comments 4 replies
-
|
nice work,please |
Beta Was this translation helpful? Give feedback.
-
|
Vote: Control Layer - Xinsir CN models |
Beta Was this translation helpful? Give feedback.
-
|
Vote: Other - Layer-Diffusion |
Beta Was this translation helpful? Give feedback.
-
|
Vote: MistoLine |
Beta Was this translation helpful? Give feedback.
-
|
Vote: Improved ToMe |
Beta Was this translation helpful? Give feedback.
-
|
Vote: Layer-Diffusion. Layer-Diffusion sounds exactly like what I would now like the most after using Krita-AI for a night. I wasn't aware of this type of plugin, I wonder if we can already manually install and use that with the Krita plugin. |
Beta Was this translation helpful? Give feedback.
-
|
We need supir! |
Beta Was this translation helpful? Give feedback.
-
|
InstantID。InstantID seems to perform better than FaceID. |
Beta Was this translation helpful? Give feedback.
-
|
yes |
Beta Was this translation helpful? Give feedback.
-
|
Vote: Other - Tooncrafter. The video shows interesting examples of multiple frames generated between two key frames. Natural and consistent. |
Beta Was this translation helpful? Give feedback.
-
|
what do you think about this. comparable to individual models or not, if the results are good then it can save a lot of space on the ssd. |
Beta Was this translation helpful? Give feedback.
-
|
Yes it sounds great. According to instruction you specify an ID depending on which mode you want, which isn't possible yet in ComfyUI as far as I know. It works without doing it, but maybe results can be better if you provide it. |
Beta Was this translation helpful? Give feedback.
-
|
Not sure if this would be an interesting candidate to implement into the plugin but any chance the AuraFlow Ai could be implemented? |
Beta Was this translation helpful? Give feedback.
-
|
@Acly: This seems pretty interesting: background removal using InSPyReNet (https://github.com/plemeri/InSPyReNet). I am already using the "transparent background" implementation (https://github.com/plemeri/transparent-background), but having this in Krita directly could be pretty interesting. Although, I admit this might be a bit different from the diffusion implementation you have done with Krita Diffusion. |
Beta Was this translation helpful? Give feedback.
-
|
https://huggingface.co/blog/quanto-diffusers good development and hopefully optimized versions of Aura Flow and ControlNet versions for it will appear soon. and then it is quite possible to consider adding it to the plugin |
Beta Was this translation helpful? Give feedback.
-
|
https://huggingface.co/black-forest-labs/FLUX.1-dev I'm sure flux support is on your radar and I look forward to building in krita with it! |
Beta Was this translation helpful? Give feedback.
-
|
Added a table for base models: Kolors, PixArt, Hunyuan-DiT, AuraFlow, Flux They come up a lot, but none of them have any kind of ecosystem or commitment from open-source community so far. Kolors is the only one with inpaint and IP-Adapter support (both coming from the model authors directly). |
Beta Was this translation helpful? Give feedback.
-
|
https://huggingface.co/2vXpSwA7/iroiro-lora/blob/main/test_controlnet2/CN-posetest_v1.safetensors 这个可以用手绘简笔骨骼图生成人物动作 |
Beta Was this translation helpful? Give feedback.
-
|
[Flux [dev]!! ControlNet for SDXL！！We need them!!! |
Beta Was this translation helpful? Give feedback.
-
|
I'd prefer to see a GGUF/unet implementation native in Krita before ControlNet on flux. But that's just my 8GB of VRAM crying in a corner... |
Beta Was this translation helpful? Give feedback.
-
|
I'd really like to see SUPIR. The current upscale implementation feels lacking |
Beta Was this translation helpful? Give feedback.
-
|
Vote: Base Models - Kolors |
Beta Was this translation helpful? Give feedback.
-
|
flux controlnet upscaler: |
Beta Was this translation helpful? Give feedback.
-
|
vote SUPIR UPSCALE |
Beta Was this translation helpful? Give feedback.
-
|
FLUX ipadapter would be a great addition. It doesn't work as strongly as the style ipadapter for SDXL, but anything is better than the way FLUX resists 'style by' prompts. https://huggingface.co/XLabs-AI/flux-ip-adapter I suggest reading this paper: https://sandner.art/prompting-art-and-design-styles-in-flux-in-forge-and-comfyui/ |
Beta Was this translation helpful? Give feedback.
-
|
Vote: IC-Light - Lighting/shading control |
Beta Was this translation helpful? Give feedback.
-
This is a list of AI models, papers and ComfyUI nodes which might be candidates for integration into the plugin.
Guide to posting feature suggestions and requests:
Important
Please do not open separate issues / discussions just to post a link!
This helps avoid clutter. Keep in mind there are many shiny new things coming out all the time, and realistically only a fraction of them will be integrated.
How to make it happen?
In order of most likely success:
Base Models
Control Layer
Performance
Other
Beta Was this translation helpful? Give feedback.
All reactions