-
Version 1.0.0 beta 24
Python 3.11.8
only web-frontend (qt5-gui does not work, see #140)
mode: LCM-openVINO
model: "rupeshs/sd-turbo-openvino"
inference steps: 6
image width, height: 512, 512
nu…
-
Could you please make a step-by-step guide on how to install FASTSDCPU on WINDOWS from scratch
-
The README states the minimum Python version is 3.8, but it seems 3.9 is required because of `networkx`:
```
Starting FastSD CPU env installation...
Found python3 command
Python version : 3.8.10…
-
-
### Describe the bug
The `load_lora_weights ` is not working offline.
### Reproduction
Sample code to reproduce this issue. (Turn off the internet)
```py
from diffusers import Diffus…
-
```
Traceback (most recent call last):
File "/root/fastsdcpu/src/app.py", line 169, in
from frontend.webui.ui import start_webui
File "/root/fastsdcpu/src/frontend/webui/ui.py", line 6, i…
-
It's possible to adapt pipeline and convert weights to openvino format with little hackaround.
Current missing feature is implementation of timestep_cond input in compiled unet, which breakes guid…
-
Well, today I tried to use `openvinotoolkit/stable-diffusion-webui` because I want to run the model `ChilloutMix-Ni-pruned-fp16` and `DreamShaper XL`, which fastdcpu haven't supported. However, I find…
-
I followed the guidance and downloaded the `SDXL Turbo OpenVINO int8 ` and `TAESDXL OpenVINO` for offline use. However, I encountered with an error. Here are my outputs:
```
Starting FastSD CPU plea…
-
Do you have any plans on running he project for Android too? Maybe you already have such a manual?
Thank you for understanding and for such a great job you do!