-
I'm having trouble inference with diffusers.
When I convert fine-tuned ckpt to diffusers and then inference, I'm getting a completely different image than ckpt.
The inference code is taken from d…
-
### Describe the bug
argument of type 'NoneType' is not iterable
checkpoint = os.path.join(hub_dir, "checkpoints/StableDiffusionXL/model/ud_sdxl-动漫二次元.safetensors")
pipe = StableDiffusionCon…
-
# 新增模型Hotshot-XL
## 任务描述
### 任务背景
- Hotshot-XL是一个text-to-GIF模型, 可以与SDXL、ControlNet兼容工作。
### 完成步骤
1. 模型、代码已经开源。
2. 根据开源代码进行网络结构转换,训练、推理,[代码链接](https://github.com/hotshotco/Hotshot-XL)。
3. 前向推理…
-
### Checklist
- [X] The issue exists after disabling all extensions
- [X] The issue exists on a clean installation of webui
- [ ] The issue is caused by an extension, but I believe it is caused b…
-
### Is there an existing issue for this?
- [X] I have searched the existing issues and checked the recent builds/commits
### What happened?
It worked well at first two times but it got error like a…
-
Hi guys I am using this [colab notebook by pedrogengo](https://colab.research.google.com/drive/1iSFDpRBKEWr2HLlz243rbym3J2X95kcy?usp=sharing)
For unknown reasons, I keep getting **nan** loss during…
-
For those who like command line interfaces, and dislike gradio and it's timeouts, and have a M1 CPU, which does things slowly, you can use this code example:
```
import argparse
import torch
imp…
-
**Describe the bug**
I can use ComViz load_model_from_config function , add safetensors load , it's work , but when I use deepspeed get error _pickle.UnpicklingError: invalid load key, '\xec'
…
-
With current version in colab I´m getting following error:
```
Training the UNet...
...
Traceback (most recent call last):
File "/content/diffusers/examples/dreambooth/train_dreambooth.py", lin…
-
Another question: `convert_lora_safetensor_to_diffusers.py` converts safetensors to diffusers format. After I trained LoRA model, I have the following in the output folder and checkpoint subfolder:
…