Adds logic to `DiskLatentsStorage.start()` to empty the latents folder on startup.
Adds start and stop methods to `ForwardCacheLatentsStorage`. This is required for `DiskLatentsStorage.start()` to be called, due to how this particular service breaks the direct DI pattern, wrapping the underlying storage with a cache.
## What type of PR is this? (check all applicable)
- [ ] Refactor
- [ ] Feature
- [X] Bug Fix
- [ ] Optimization
- [ ] Documentation Update
- [ ] Community Node Submission
## Have you discussed this change with the InvokeAI team?
- [X] Yes
- [ ] No, because:
## Have you updated all relevant documentation?
- [X] Yes
- [ ] No
## Description
This adds support for at least some of the SDXL embeddings currently
available on Civitai. The embeddings I have tested include:
- https://civitai.com/models/154898/marblingtixl?modelVersionId=173668
- https://civitai.com/models/148131?modelVersionId=167640
-
https://civitai.com/models/123485/hannah-ferguson-or-sdxl-or-comfyui-only-or-embedding?modelVersionId=134674
(said to be "comfyui only")
-
https://civitai.com/models/185938/kendall-jenner-sdxl-embedding?modelVersionId=208785
I am _not entirely sure_ that I have implemented support in the most
elegant way. The issue is that these embeddings have two weight tensors,
`clip_g` and `clip_l`, which correspond to `text_encoder` and
`text_encoder_2` in the main model. When the patcher calls the
ModelPatcher's `apply_ti()` method, I simply check the dimensions of the
incoming text encoder and choose the weights that match the dimensions
of the encoder.
While writing this, I also ran into a possible issue with the Compel
library's `get_pooled_embeddings()` call. It pads the input token list
to the model's max token length and then calls the TI manager to add the
additional tokens from the embedding. However, this ends up making the
input token list longer than the max length, and CLIPTextEncoder crashes
with a tensor size mismatch. I worked around this behavior by making the
TI manager's `expand_textual_inversion_token_ids_if_necessary()` method
remove the excess pads at the end of the token list.
Also note that I have made similar changes to `apply_ti()` in the
ONNXModelPatcher, but haven't tested them yet.
## Related Tickets & Documents
<!--
For pull requests that relate or close an issue, please include them
below.
For example having the text: "closes #1234" would connect the current
pull
request to issue 1234. And when we merge the pull request, Github will
automatically close the issue.
-->
- Related Issue #
- Closes#4401
## QA Instructions, Screenshots, Recordings
<!--
Please provide steps on how to test changes, any hardware or
software specifications as well as any other pertinent information.
-->
## Added/updated tests?
- [ ] Yes
- [X] No : We need to create tests for model patching...
## [optional] Are there any post deployment tasks we need to perform?
IndexedDB has a much larger storage limit than LocalStorage, and is widely supported.
Implemented as a custom storage driver for `redux-remember` via `idb-keyval`. `idb-keyval` is a simple wrapper for IndexedDB that allows it to be used easily as a key-value store.
The logic to clear persisted storage has been updated throughout the app.
- Reset init image, control adapter images, and node image fields when their selected image fails to load
- Only do this if the app is connected via socket (this indicates that the image is "really" gone, and there isn't just a transient network issue)
It's possible for image parameters/nodes/states to have reference a deleted image. For example, a resize image node might have an image set on it, and the workflow saved. The workflow contains a hard reference to that image.
The image is deleted and the workflow loaded again later. The deleted image is still in that workflow, but the app doesn't detect that. The result is that the workflow/graph appears to be valid, but will fail on invoke.
This creates a really confusing user experience, where when somebody shares a workflow with an image baked into it, and another person opens it, everything *looks* ok, but the workflow fails with a mysterious error about a missing image.
The problem affects node images, control adapter images and the img2img init image. Resetting the image when it fails to load *and* socket is connected resolves this in a simple way.
The problem also affects canvas images, but we have handle that by displaying an error fallback image, so no change is made there.
Closes#5121
- Parse `anyOf` for enums (present when they are optional)
- Consolidate `FieldTypeParseError` and `UnsupportedFieldTypeError` into `FieldParseError` (there was no difference in handling and it simplifies things a bit)
* add centerpadcrop node
- Allows users to add padding to or crop images from the center
- Also outputs a white mask with the dimensions of the output image for use with outpainting
* add CenterPadCrop to NODES.md
Updates NODES.md with CenterPadCrop entry.
* remove mask & output class
- Remove "ImageMaskOutput" where both image and mask are output
- Remove ability to output mask from node
---------
Co-authored-by: psychedelicious <4822129+psychedelicious@users.noreply.github.com>
