`torch` wasn't seeing the environment variable. I suspect this is because it was imported before the variable was set, so was running with a different environment.
Many `torch` ops are supported on MPS so this wasn't noticed immediately, but some samplers like k_dpm_2 still use unsupported operations and need this fallback.
* remove non maintained Dockerfile
* adapt Docker related files to latest changes
- also build the frontend when building the image
- skip user response if INVOKE_MODEL_RECONFIGURE is set
- split INVOKE_MODEL_RECONFIGURE to support more than one argument
* rename `docker-build` dir to `docker`
* update build-container.yml
- rename image to invokeai
- add cpu flavor
- add metadata to build summary
- enable caching
- remove build-cloud-img.yml
* fix yarn cache path, link copyjob
Crashes would occur in the invokeai-configure script if no HF token
was found in cache and the user declines to provide one when prompted.
The reason appears to be that on Linux systems getpass_asterisk()
raises an EOFError when no input is provided
On windows10, getpass_asterisk() does not raise the EOFError, but
returns an empty string instead. This patch detects this and raises
the exception so that the control logic is preserved.
if reinstalling over an existing installation where the .venv was
created with symlinks to system python instead of copies of the python
executable, the installer would raise a `SameFileError`, because it
would attempt to copy Python over itself. This fixes the issue.
Copying the executable is still preferred for new environments, because
this guarantees the stable Python version.
- fixes bug in finding the source of the configs dir;
- updates the docs for manual install to clarify the preference to
keeping the `.venv` inside the runtime dir, and the caveat/extra steps
required if done otherwise
if reinstalling over an existing installation where the .venv
was created with symlinks to system python instead of copies
of the python executable, the installer would raise a
SameFileError, because it would attempt to copy Python over
itself. This fixes the issue.
- Added modest adaptive behavior; if the screen is wide enough the three
checklists of models will be arranged in a horizontal row.
- Added color support
## Summary
This PR rewrites the core of the installer in Python for cross-platform
compatibility. Filesystem path manipulation, platform/arch decisions and
various edge cases are handled in a more convenient fashion. The
original `install.bat.in`/`install.sh.in` scripts are kept as
entrypoints for their respective OSs, but only serve as thin wrappers to
the Python module.
In addition, it:
- builds and **packages the .whl with the installer**, so that
downloading a versioned installer will guarantee installation of the
same version of the application.
- updates shell entrypoints:
- new commands are `invokeai`, `invokeai-configure`, `invokeai-ti`,
`invokeai-merge`.
- these commands will be available in the activated `.venv` or via the
launch scripts
- `invoke.py` and `configure_invokeai.py` scripts are deprecated but
kept around for backwards compatibility and keeping users' surprise to a
minimum.
- introduces a new `ldm/invoke/config` package and moves the
`configure_invokeai` script into it. Similarly, movers Textual Inversion
script and TUI to `ldm/invoke/training`.
- moves the `configs` directory into the `ldm/invoke/config` package for
easy distribution.
- updates documentation to reflect all of the above changes
- fixes a failing test
- reduces wheel size to 3MB (from 27MB) by excluding unnecessary image
files under `assets`
⚠️ self-updating functionality and ability to install arbitrary
versions are still WIP. For now we can recommend downloading and running
the installer for a specific version as desired.
## Testing the source install
From the cloned source, check out this branch, and:
`$ python3 installer/main.py --root <path_to_destination>`
Also try:
`$ python3 installer/main.py ` - will prompt for paths
`$ python3 installer/main.py --yes` - will not prompt for any input
- try to combine the `--yes` and `--root` options
- try to install in destinations with "quirky" paths, such as paths
containing spaces in the directory name, etc.
## Testing the packaged install ("Automated Installer"):
Download the
[InvokeAI-installer-v2.3.0+a0.zip](https://github.com/invoke-ai/InvokeAI/files/10533913/InvokeAI-installer-v2.3.0%2Ba0.zip)
file, unzip it, and run the install script for your platform (preferably
in a terminal window)
OR make your own: from the cloned source, check out this branch, and:
```
cd installer
./create_installer.sh
# (do NOT tag/push when prompted! just say "no")
```
This will create the installation media:
`InvokeAI-installer-v2.3.0+a0.zip`. The installer is now
*platform-agnostic* - meaning, both Windows and *nix install resources
are packaged together.
Copy it somewhere as if it had been downloaded from the internet. Unzip
the file, enter the created `InvokeAI-Installer` directory, and run
`install.sh` or `install.bat` as applicable your platform.
⚠️ NOTE!!! `install.sh` accepts the same arguments as are
applicable to the Python script, i.e. you can `install.sh --yes --root
....`. This is NOT yet supported by the Windows `.bat` script. Only
interactive installation is supported on Windows. (this is still a
TODO).
* refactor ckpt_to_diffuser to allow converted pipeline to remain in memory
- This idea was introduced by Damian
- Note that although I attempted to use the updated HuggingFace module
pipelines/stable_diffusion/convert_from_ckpt.py, it was unable to
convert safetensors files for reasons I didn't dig into.
- Default is to extract EMA weights.
* add --ckpt_convert option to load legacy ckpt files as diffusers models
- not quite working - I'm getting artifacts and glitches in the
converted diffuser models
- leave as draft for time being
* do not include safety checker in converted files
* add ability to control which vae is used
API now allows the caller to pass an external VAE model to the
checkpoint conversion process. In this way, if an external VAE is
specified in the checkpoint's config stanza, this VAE will be used
when constructing the diffusers model.
Tested with both regular and inpainting 1.X models.
Not tested with SD 2.X models!
---------
Co-authored-by: Jonathan <34005131+JPPhoto@users.noreply.github.com>
Co-authored-by: Damian Stewart <null@damianstewart.com>
