I updated the paths to make all paths work
Now paths in ComfyUI are working
Added:
- `input`
- `models/configs`
- `models/gligen`
- `models/diffusers`
- `models/hypernetworks`
Was broken:
- `custom_nodes`
- `models/clip_vision`
- `models/clip`
---------
Co-authored-by: LEv145 <you@example.com>
Co-authored-by: AbdBarho <ka70911@gmail.com>
`jq` merge direction in this case is right to left so if the user had
set up custom paths it would replace them with the default ones.
This PR switches the direction to use the defaults as fallback instead
of overwriting user settings.
----
Didn't want to create an issue for the tiny change.
Thanks for your work on the repo, it saved me a lot of time, 👍
---------
Co-authored-by: AbdBarho <ka70911@gmail.com>
As discussed in Discussion
[#367](https://github.com/AbdBarho/stable-diffusion-webui-docker/discussions/367),
this adds support for the newer ComfyUI. I forked the fork that would
already add this but the maintainer of that fork hasn't implemented the
changes needed to properly get the output function working, which I did.
I believe everything is functional though I have not tested every single
node.
I changed the table format for the README and a few other minor things
for aesthetic reasons but if you want me to revert those, I will.
---------
Co-authored-by: Jonathan Kovacs <jkovacs-dev@users.noreply.github.com>
Co-authored-by: AbdBarho <ka70911@gmail.com>
## Justification
Closes issue #352
This update makes the Dockerfiles OCI compliant, making it easier to use
Buildah or other image building techniques that require it
## Implementation
This changes a few things, listed below:
* auto: Download container is switched to alpine. The `git` container
specified the `/git` directory as a volume. As such, all the files under
`/git` would be lost after each script invoke. Alpine is used later in
the build process anyway, so it shouldn't be any extra cost to switch to
it
* auto: "New" clone.sh script is copied into the container, which is
basically just the previous clone script that was embedded in the
Dockerfile.
* all: `<<EOF` heredoc styles have been switched to `&& \`
* all: I added NVIDIA_DRIVER_CAPABILITIES and NVIDIA_VISIBLE_DEVICES to
expose my Nvidia card. This is most likely a selinux/podman problem, but
shouldn't change anything with docker to add it.
* docker-compose: I added selinux labeling. I tested this with real
docker (not just podman!) and it seems to work fine. Though I suggest
you try it too.
## Testing
Locally builds with buildah.
Note: for caching to work properly, you still need to replace
`/root/.cache/pip` with `/root/.cache/pip,Z` on selinux systems.
Note: I was having some trouble running invoke. Thought it was this PR,
but it's a known issue. See
https://github.com/invoke-ai/InvokeAI/issues/3182
---------
Co-authored-by: AbdBarho <ka70911@gmail.com>
Upon enabling the ControlNet addon from
https://github.com/AbdBarho/stable-diffusion-webui-docker/pull/385 one
might want to use the `openpose` preprocessors. Those are downloaded by
the addon the first time they are used. Without proper mounts those
networks will be downloaded on usage after each container start.
This PR enables those mounts to reduce data traffic.
The ControlNet addon
[sd-webui-controlnet](https://github.com/Mikubill/sd-webui-controlnet)
requires the `data/ControlNet` folder to be mounted into
`models/ControlNet`.
This PR enables said mount and adds the ControlNet folder to
`.gitignore` file.
---------
Co-authored-by: AbdBarho <ka70911@gmail.com>
Rename the UIs in docker compose to their new names
Changes folder names
Changes output folder structure
Closes issue #263
Adds `sygil-sl` instead of docker compose flag.
Builds on top of #150
Thanks to @pirahtays
Co-authored-by: Imaginator <mriegel@gmail.com>
Co-authored-by: pirahtays <35934562+pirahtays@users.noreply.github.com>
- auto:
36a0ba357a
- History tab should be working, closes#138
- hlky:
bd57d22f2e
- experimental support for the streamlit UI, use the env var
`USE_STREAMLIT`, see the `docker-compose.yml` file, closes#105
- Don't create issues if it fails, it is still very early in dev.
### Update versions
- auto:
03d62538ae
- History Tab IS NOT WORKING YET! #138
- hlky:
fd51bab1ec
- lstein:
fe2a2cfc8bCloses#102 the config file has been moved to `data/config/auto`
1. Update `docker-compose.yml`: use `PRELOAD=true` for lstein. See the
[note](94bad8555a/docs/installation/INSTALL_LINUX.md?plain=1#L60)
in the installation guide of InvokeAI.
3. Fix `services/lstein/mount.sh`: change `CodeFormer` to `Codeformer`.
4. Update `services/lstein/mount.sh`: avoid having to re-download the
`${PWD}/gfpgan/weights/...` every time the container is started.
- auto:
050a6a798c
- Now uses python 3.10
- requires a complete re-install
- Image is now smaller (5.7GB vs 9.8GB)
- hlky:
fe6e72fde7
- lstein:
31869885d9
- img2img now works
- auto:
2995107fa2
- More samplers
- Textual inversion training
- hlky:
1a9c053cb7
- Build times are SLOW
- lstein:
4f247a3672
- Prepare for 2.0 release
- very cool new UI!
### Update versions
- auto:
3f417566b0
### Breaking changes:
* renamed `automatic-1111` service to `auto`
* the `cache` folder is now deprecated, replaced with `data` (see
migration guide below)
* `embeddings` folder has been moved to `data/embeddings`
* use GFPGAN 1.4
### Migration Guide
Note: in theory, running the command
```
docker compose --profile download up --build
```
is all you need to use the new version, however, this means you will
also have to download everything again. A new script is available under
`scripts/migratev1tov2.sh` that will copy models to the new structure
and should get you most of the way, run
```bash
./scripts/migratev1tov2.sh
```
or you can manually inspect the script and copy the files
After that, run
```
docker compose --profile download up --build
```
to validate everything.
