Releases: ashleykleynhans/stable-diffusion-docker
1.6.4
ce280bf
ashleykleynhans
Ashley Kleynhans
- Bump Kohya_ss to v21.8.5.
- Added option to download SDXL models.
- Download SDXL styles.
1.6.1
87b7583
ashleykleynhans
Ashley Kleynhans
- Bumped Automatic1111 Web UI to version 1.5.1.
1.6.0
edacae8
ashleykleynhans
Ashley Kleynhans
- Bumped Automatic1111 Web UI to version 1.5.0.
- Now with SDXL support.
1.5.3
3132f0d
ashleykleynhans
Ashley Kleynhans
- Added nginx reverse proxy.
- Added HTML for HTTP 502 errors to inform user that port is not up yet.
- Dropped gradio version that was causing the Stable Diffusion Web UI to be unresponsive.
- Bumped Kohya_ss to v21.8.2.
- Bumped Dreambooth to commit c2a5617c587b812b5a408143ddfb18fc49234edf.
- Added caching of Stable Diffusion 1.5 model for faster start-up time.
1.5.1
e39fedb
ashleykleynhans
Ashley Kleynhans
Automatic1111 Stable Diffusion WebUI with ControlNet, Deforum, Dreambooth, roop extensions + Kohya SS
Version 1.5.1
Included in this Template
- Ubuntu 22.04 LTS
- CUDA 11.8
- Python 3.10.6
- Automatic1111 Stable Diffusion Web UI 1.4.0
- Dreambooth extension 1.0.14
- Deforum extension
- ControlNet extension v1.1.229
- Additional networks extension
- Locon extension
- roop extension 0.0.2
- Kohya_ss v21.7.16
- Torch 2.0.1
- xformers 0.0.20
- v1-5-pruned.safetensors
- vae-ft-mse-840000-ema-pruned.safetensors
Ports
| Port | Description |
|---|---|
| 3000 | Stable Diffusion Web UI |
| 3010 | Kohya_ss |
| 6006 | Tensorboard |
| 8888 | Jupyter Lab |
Environment Variables
| Variable | Description | Default |
|---|---|---|
| JUPYTER_PASSWORD | Password for Jupyter Lab | Jup1t3R! |
| DISABLE_AUTOLAUNCH | Disable Web UIs from launching automatically | enabled |
| ENABLE_TENSORBOARD | Enables Tensorboard on port 6006 | disabled |
Logs
Stable Diffusion Web UI and Kohya SS both create log
files, and you can tail the log files instead of
killing the services to view the logs
| Application | Log file |
|---|---|
| Stable Diffusion Web UI | /workspace/logs/webui.log |
| Kohya SS | /workspace/logs/kohya_ss.log |
For example:
tail -f /workspace/logs/webui.logJupyter Lab
If you wish to use the Jupyter lab, you must set
the JUPYTER_PASSWORD environment variable in the
Template Overrides configuration when deploying
your pod.
General
Note that this does not work out of the box with
encrypted volumes!
This is a custom packaged template for Stable Diffusion
using the Automatic1111 Web UI, as well as the Dreambooth,
Deforum, ControlNet and roop extension repos.
It also contains the Kohya_ss Web UI.
I do not maintain the code for any of these repos,
I just package everything together so that it is
easier for you to use.
If you need help with settings, etc. You can feel free
to ask me, but just keep in mind that I am not an expert
at Stable Diffusion! I'll try my best to help, but the
RunPod community or Automatic/Stable Diffusion communities
may be better at helping you.
Please wait until the GPU Utilization % is 0 before
attempting to connect. You will likely get a 502 error
before that as the pod is still getting ready to be used.
Changing launch parameters
You may be used to changing a different file for your
launch parameters. This template uses webui-user.sh,
which is located in the webui directory
(/workspace/stable-diffusion-webui) to manage the
launch flags such as --xformers. You can feel free
to edit this file, and then restart your pod via the
hamburger menu to get them to go into effect, or
alternatively just use fuser -k 3000/tcp and start
the /workspace/stable-diffusion-webui/webui.sh -f
script again.
--xformers and --api are parameters that are
frequently asked about.
Using your own models
The best ways to get your models onto your pod is
by using runpodctl or by uploading them to Google
Drive or other cloud storage and downloading them
to your pod from there.
Uploading to Google Drive
If you're done with the pod and would like to send
things to Google Drive, you can use this colab to do it
using runpodctl. You run the runpodctl either in
a web terminal (found in the pod connect menu), or
in a terminal on the desktop.
1.5.0
72fbede
ashleykleynhans
Ashley Kleynhans
Automatic1111 Stable Diffusion WebUI with Dreambooth extension + Kohya SS
Version 1.5.0
Included in this Template
- Ubuntu 22.04 LTS
- CUDA 11.8
- Python 3.10.6
- Automatic1111 Stable Diffusion Web UI 1.4.0
- Dreambooth extension 1.0.14
- Deforum extension
- ControlNet extension v1.1.227
- Kohya_ss v21.7.16
- Torch 2.0.1
- xformers 0.0.20
- v1-5-pruned.safetensors
- vae-ft-mse-840000-ema-pruned.safetensors
Ports
| Port | Description |
|---|---|
| 3000 | Stable Diffusion Web UI |
| 3010 | Kohya_ss |
| 6006 | Tensorboard |
| 8888 | Jupyter Lab |
Environment Variables
| Variable | Description | Default |
|---|---|---|
| JUPYTER_PASSWORD | Password for Jupyter Lab | Jup1t3R! |
| DISABLE_AUTOLAUNCH | Disable Web UIs from launching automatically | enabled |
| ENABLE_TENSORBOARD | Enables Tensorboard on port 6006 | disabled |
Logs
Stable Diffusion Web UI and Kohya SS both create log
files, and you can tail the log files instead of
killing the services to view the logs
| Application | Log file |
|---|---|
| Stable Diffusion Web UI | /workspace/logs/webui.log |
| Kohya SS | /workspace/logs/kohya_ss.log |
For example:
tail -f /workspace/logs/webui.logJupyter Lab
If you wish to use the Jupyter lab, you must set
the JUPYTER_PASSWORD environment variable in the
Template Overrides configuration when deploying
your pod.
General
Note that this does not work out of the box with
encrypted volumes!
This is a custom packaged template for Stable Diffusion
using the Automatic1111 Web UI, as well as the Dreambooth,
Deforum and ControlNet extension repos.
It also contains the Kohya_ss Web UI.
I do not maintain the code for any of these repos,
I just package everything together so that it is
easier for you to use.
If you need help with settings, etc. You can feel free
to ask me, but just keep in mind that I am not an expert
at Stable Diffusion! I'll try my best to help, but the
RunPod community or Automatic/Stable Diffusion communities
may be better at helping you.
Please wait until the GPU Utilization % is 0 before
attempting to connect. You will likely get a 502 error
before that as the pod is still getting ready to be used.
Changing launch parameters
You may be used to changing a different file for your
launch parameters. This template uses webui-user.sh,
which is located in the webui directory
(/workspace/stable-diffusion-webui) to manage the
launch flags such as --xformers. You can feel free
to edit this file, and then restart your pod via the
hamburger menu to get them to go into effect, or
alternatively just use fuser -k 3000/tcp and start
the /workspace/stable-diffusion-webui/webui.sh -f
script again.
--xformers and --api are parameters that are
frequently asked about.
Using your own models
The best ways to get your models onto your pod is
by using runpodctl or by uploading them to Google
Drive or other cloud storage and downloading them
to your pod from there.
Uploading to Google Drive
If you're done with the pod and would like to send
things to Google Drive, you can use this colab to do it
using runpodctl. You run the runpodctl either in
a web terminal (found in the pod connect menu), or
in a terminal on the desktop.
1.4.2
40108d8
ashleykleynhans
Ashley Kleynhans
Included in this Image
- Ubuntu 22.04 LTS
- CUDA 11.8
- Python 3.10.6
- Automatic1111 Stable Diffusion Web UI 1.3.0
- Dreambooth extension (dev branch)
- Deforum extension
- ControlNet extension v1.1.206
- Torch 2.0.1
- xformers 0.0.20 (disabled by default)
- v1-5-pruned.safetensors
- vae-ft-mse-840000-ema-pruned.safetensors
Jupyter Lab
If you wish to use the Jupyter lab, you must set
the JUPYTER_PASSWORD environment variable in the
Template Overrides configuration when deploying
your pod.
xformers
This template intentionally disables xformers by default
due to the fact that it reduces training quality when
training Dreambooth.
If you require xformers, you can add it back to the
COMMANDLINE_ARGS within webui-user.sh. It is
currently commented out, you can simply uncomment it
and restart the Stable Diffusion Web UI.
General
Note that this does not work out of the box with
encrypted volumes!
This is a custom packaged template for Stable Diffusion
using the Automatic1111 Web UI, as well as the Dreambooth,
Deforum and ControlNet extension repos.
I do not maintain the code for any of these repos,
I just package everything together so that it is
easier for you to use.
If you need help with settings, etc. You can feel free
to ask me, but just keep in mind that I am not an expert
at Stable Diffusion! I'll try my best to help, but the
RunPod community or Automatic/Stable Diffusion communities
may be better at helping you.
Please wait until the GPU Utilization % is 0 before
attempting to connect. You will likely get a 502 error
before that as the pod is still getting ready to be used.
Changing launch parameters
You may be used to changing a different file for your
launch parameters. This template uses launcher.py,
which is located in the webui directory
(/workspace/stable-diffusion-webui) to manage the
launch flags such as --xformers. You can feel free
to edit this file, and then restart your pod via the
hamburger menu to get them to go into effect, or
alternatively just use fuser -k 3000/tcp and start
the launcher.py script again.
--xformers and --api are parameters that are
frequently asked about.
Using your own models
The best ways to get your models onto your pod is
by using runpodctl or by uploading them to Google
Drive or other cloud storage and downloading them
to your pod from there.
Uploading to Google Drive
If you're done with the pod and would like to send
things to Google Drive, you can use this colab to do it
using runpodctl. You run the runpodctl either in
a web terminal (found in the pod connect menu), or
in a terminal on the desktop.