-
i noticed that when generating images using variant=anythingv3 the colors of the generated picture are washed out
this is most likely due to the vae not beeing used correctly i would assume.
this p…
-
通过ControlNet的原模型和Auto1111的插件均仔细体验了ControlNet的效果,目前主要使用的方式有Canny和Scribble和 Fake Scribble 三种方式,使用了作者提供的原模型,以及AnythingV3所混合的模型(有模型混合版,也有通过Auto1111实现的非混合直接调用AnythingV3的方式)
**所放置的测试图片内容如下:**
![xi…
-
### Is there an existing issue for this?
- [X] I have searched the existing issues and checked the recent builds/commits
### What happened?
after update + extension update error on some models
use…
-
### What happened?
I've noticed that on a LoRA I've made the downloaded files do not have the same name as the one's that I've uploaded. For example, I have two versions of a LoRA that I made for com…
-
I keep getting the following error which I believe is affecting other network models from loading
```
size = du_state_dict[list(du_state_dict.keys())[0]].size() # if conv2d size is like […
-
root@3776108399f0:/mmagic# python demo/gradio_controlnet_animation.py
Traceback (most recent call last):
File "/opt/conda/lib/python3.8/site-packages/huggingface_hub/utils/_errors.py", line 259, i…
-
Protogen seems to be new, really good, and easy-to-use. It also has a style which feels like it would work well with games. I think Protogen should be experimented with.
Deets
- Civitai: [Protogen…
-
Hello there,
I saw that certain ckpt models asks users to have the corresponding `vae.pt` file to be used in parallel for some corrections and better results. Only examples that come to mind are fr…
-
Hello,
when i try to run shark for my 7900xt, but the lastest stabel version has the same issue
shark_tank local cache is located at C:\Users\Cangjin\.local/shark_tank/ . You may change this by sett…
-
# Train a LoRA model for an anime character | BoHolder's site: crafts, blogs etc
From data collecting to training
[https://boholder.github.io/en-us/blogs/train-a-lora-model-for-an-anime-character/](…