* Fix to LoadImage Node for #3416 HDR images loading additional smaller images.
Added a blocking if statement in the ImageSequence.Iterator that checks if subsequent images after the first match dimensionally, and prevent them from being appended to output_images if they do not match.
This does not fix or change current behavior for PIL 10.2.0 where the images are loaded at the same size, but it does for 10.3.0 where they are loaded at their correct smaller sizes.
* added list of excluded formats that should return 1 image
added an explicit check for the image format so that additional formats can be added to the list that have problematic behavior.
* Update node_helpers.py to use generic pillow wrapper to resolve multiple meta-data related issues.
replaced open_image function with a generic pillow function that takes Pil functions as a dependency injection and applies the ImageFile.LOAD_TRUNCATED_IMAGES try except fix to them.
This provides an extensible function to handle related errors that can wrap offending functions when discovered without the need to repeat code.
* Update a few Pil functions to use node_helpers.pillow wrapper
Update a Pil function calls in a few locations to use the generic node_helpers.pillow wrapper that takes the function as a dependency injection and uses the try except method with ImageFIle.LOAD_TRUNCATED_IMAGES solution
* Corrected comment in issue #s fixed.
* Update node_helpers.py to remove import of Image from PIL
import of Image is no longer required as functions are Injected
* Fix issue with how PIL loads small PNG files nodes.py
Added flag to prevent ValueError: Decompressed Data Too Large
when loading PNG images with large meta data such as large embedded color profiles
* Update LoadImage node to fix error when loading PNG's in nodes.py
Fixed Value Error: Decompressed Data Too Large thrown by PIL when attempting to opening PNG files with large embedded ICC colorspaces by setting the follow flag to true when loading png images: ImageFile.LOAD_TRUNCATED_IMAGES = True
* Update node_helpers.py to include open_image helper function
open_image includes try except to catch Pillow Value Errors that occur when large ICC profiles are embedded in images.
* Update LoadImage node to use open_image helper function inplace of Image.open
open_image helper function in node_helpers.py fixes a Pillow error when attempting to open images with large embedded ICC profiles by adding an exception handler to load the image with truncated meta data if regular loading is not possible.
This sampler is an LCM sampler that upscales the latent during sampling.
It can be used to generate at a higher resolution with an LCM model very
quickly.
To try it use it with a basic 5 step LCM workflow with scale_ratio 1.5 or
2.0
* Implement Differential Diffusion
* Cleanup.
* Fix.
* Masks should be applied at full strength.
* Fix colors.
* Register the node.
* Cleaner code.
* Fix issue with getting unipc sampler.
* Adjust thresholds.
* Switch to linear thresholds.
* Only calculate nearest_idx on valid thresholds.
* First SAG test
* need to put extra options on the model instead of patcher
* no errors and results seem not-broken
* Use @ashen-uncensored formula, which works better!!!
* Fix a crash when using weird resolutions. Remove an unnecessary UNet call
* Improve comments, optimize memory in blur routine
* SAG works with sampler_cfg_function
The img2vid model is conditioned on clip vision output only which means
there's no CLIP model which is why I added a ImageOnlyCheckpointLoader to
load it. Note that the unClipCheckpointLoader can also load it because it
also has a CLIP_VISION output.
SDV_img2vid_Conditioning is the node used to pass the right conditioning
to the img2vid model.
VideoLinearCFGGuidance applies a linearly decreasing CFG scale to each
video frame from the cfg set in the sampler node to min_cfg.
SDV_img2vid_Conditioning can be found in conditioning->video_models
ImageOnlyCheckpointLoader can be found in loaders->video_models
VideoLinearCFGGuidance can be found in sampling->video_models
This doesn't affect how percentages behave in the frontend but breaks
things if you relied on them in the backend.
percent_to_sigma goes from 0 to 1.0 instead of 1.0 to 0 for less confusion.
Make percent 0 return an extremely large sigma and percent 1.0 return a
zero one to fix imprecision.
This node takes a list of sigmas and a sampler object as input.
This lets people easily implement custom schedulers and samplers as nodes.
More nodes will be added to it in the future.
The created checkpoints contain workflow metadata that can be loaded by
dragging them on top of the UI or loading them with the "Load" button.
Checkpoints will be saved in fp16 or fp32 depending on the format ComfyUI
is using for inference on your hardware. To force fp32 use: --force-fp32
Anything that patches the model weights like merging or loras will be
saved.
The output directory is currently set to: output/checkpoints but that might
change in the future.
* wip
* latent dir
* fix
* fix
* now working
* mark todo
* remove server.py changes to separate PRt
---------
Co-authored-by: Lt.Dr.Data <lt.dr.data@gmail.com>
* allow nodes to map over lists
* make work with IS_CHANGED and VALIDATE_INPUTS
* give list outputs distinct socket shape
* add rebatch node
* add batch index logic
* add repeat latent batch
* deal with noise mask edge cases in latentfrombatch
* Add clipspace feature.
* feat: copy content to clipspace
* feat: paste content from clipspace
Extend validation to allow for validating annotated_path in addition to other parameters.
Add support for annotated_filepath in folder_paths function.
Generalize the '/upload/image' API to allow for uploading images to the 'input', 'temp', or 'output' directories.
* rename contentClipboard -> clipspace
* Do deep copy for imgs on copy to clipspace.
* mask painting on clipspace
* add original_imgs into clipspace
* Preserve the original image when 'imgs' are modified
* robust patch & refactoring folder_paths about annotated_filepath
* wip
* Only show the Paste menu if the ComfyApp.clipspace is not empty
* clipspace feature added
maskeditor feature added
* instant refresh on paste
force triggering 'changed' on paste action
* enhance mask painting
smooth drawing
add brush_size +/- button
* robust patch
use mouseup event
* robust patch
again...
* subfolder fix on paste logic
attach subfolder if subfolder isn't empty
* event listener patch
add ], [ key event for brush size
remove listener on close
* Fix button positioning issue related to window height.
Change brush size from button to slider.
* clean commit
* clean code
* various bug fixes
* paste action
- prevent opening upload popup
- ensure rendering after widget_value update
* view api update
- support annotated_filepath
* maskeditor layout
- prevent covering button by hidden div
* remove dbg message
* Add cursor functionality to display brush size
* refactor: Replace brush preview feature with missionfloyd implementation
* missionfloyd implementation
* hiding brush preview off the canvas
* change brush size on wheel event
* keyup -> keydown event
* Update web/extensions/core/maskeditor.js
Co-authored-by: missionfloyd <missionfloyd@users.noreply.github.com>
* Add support for channel-specific image data retrieval in /view API to fix alpha mask loading issue
When loading an image with an alpha mask in JavaScript canvas, there is an issue where the alpha and RGB channels are premultiplied. To avoid reliance on JavaScript canvas, I added support for channel-specific image data retrieval in the "/view" API. This allows us to retrieve data for each channel separately and fix the alpha mask loading issue. The changes have been committed to the repository.
* Enable brush preview for key and slider events
* optimize
* preview fix
* robust patch
* fix copy (clipspace) action
imgs[0] copy -> whole imgs copy
* support batch images on clipspace, maskeditor
* copy/paste bug fixes for batch images
enhance selector preview on clipspace menu
add img_paste_mode option into clipspace menu
* crash fix
* print message if clipspace content cannot editable
* Update web/extensions/core/maskeditor.js
Co-authored-by: missionfloyd <missionfloyd@users.noreply.github.com>
* make default img_paste_mode to 'selected'
refactor space -> tab
* save clipspace files to input/clipspace instead of temp
* show "clipspace/filename.png" instead of 'filename.png [clipspace]' in LoadImage/LoadImageMask
* refresh fix related to FILE_COMBO
* Update web/extensions/core/maskeditor.js
Co-authored-by: missionfloyd <missionfloyd@users.noreply.github.com>
* Update web/extensions/core/maskeditor.js
Co-authored-by: missionfloyd <missionfloyd@users.noreply.github.com>
* adjust margin based on missionfloyd impelements
* mouse event -> pointer event
* pen, touch, mouse drawing patched and tested
* Update web/extensions/core/maskeditor.js
Co-authored-by: missionfloyd <missionfloyd@users.noreply.github.com>
* add comment about touch event.
---------
Co-authored-by: Lt.Dr.Data <lt.dr.data@gmail.com>
Co-authored-by: missionfloyd <missionfloyd@users.noreply.github.com>
The mask is dilated by grow_mask_by pixels after being applied to the pixel
space image. This helps reduce seams caused by inpainting. Higher value
means less seams.
* first commit
* fixed a bunch of things missing in initial commit.
* parameters renamed for clarity
* renamed node, attempted update cond list
* to_strength removed, it is now normalized
* removed comments and prints. Attempted to apply to every cond in list again but no luck
* fixed repeating frames after batch using deepcopy
* Revert "fixed repeating frames after batch using deepcopy"
This reverts commit 1086d6a0e1.
* Rewrite addWeighted to use torch.mul iteratively.
---------
Co-authored-by: City <125218114+city96@users.noreply.github.com>
By default, when applying a mask to a condition, the entire image will
still be used for sampling. The new "set_area_to_bounds" option on the
node will allow the user to automatically limit conditioning to the
bounds of the mask.
I've also removed the dependency on torchvision for calculating bounding
boxes. I've taken the opportunity to fix some frustrating details in the
other version:
1. An all-0 mask will no longer cause an error
2. Indices are returned as integers instead of floats so they can be
used to index into tensors.
* Add clipspace feature.
* feat: copy content to clipspace
* feat: paste content from clipspace
Extend validation to allow for validating annotated_path in addition to other parameters.
Add support for annotated_filepath in folder_paths function.
Generalize the '/upload/image' API to allow for uploading images to the 'input', 'temp', or 'output' directories.
* rename contentClipboard -> clipspace
* Do deep copy for imgs on copy to clipspace.
* add original_imgs into clipspace
* Preserve the original image when 'imgs' are modified
* robust patch & refactoring folder_paths about annotated_filepath
* Only show the Paste menu if the ComfyApp.clipspace is not empty
* instant refresh on paste
force triggering 'changed' on paste action
* subfolder fix on paste logic
attach subfolder if subfolder isn't empty
---------
Co-authored-by: Lt.Dr.Data <lt.dr.data@gmail.com>
