Added FileBackedDict, which provides a dictionary-like interface with file-backed persistence. Supports nested dictionaries and ensures data is stored and retrieved from disk.
video.py:
Implemented audio resampling in CombineAudioVideoFiles, including video and audio resampling logic.
Refined create_video_from_images to use ffmpeg for combining image sequences into video.
Updated CombineFoldersWithFramesToVideoFiles to handle both duration and fps-based input.
Added additional parameters for video quality and pixel format.
torchvision_nodes.py:
Centralized the imports for NODE_CLASS_MAPPINGS and NODE_DISPLAY_NAME_MAPPINGS across the module.
Adjusted default values for certain fields to improve user experience.
youtube_nodes.py & modules/youtube.py:
Extended the YouTube upload functionality to support optional thumbnail uploads.
Improved initialize_upload function to return video_id after a successful upload.
Introduced the ability to upload a thumbnail alongside the video.
web/ui_modifications.js:
Reworked the login form handling to prevent page reload and improved element management.
Added dynamic graph creation in the handleGraphCreateNew function.
web/showMarkdown.js:
Fixed the import of marked.js, using Marked as the module name to avoid potential conflicts.
modules/xttscli.py:
Increased the batch_size for text-to-speech processing from 2 to 4, enabling faster concurrent execution.
Simplified class imports and added fallback mechanisms for missing modules.
Enhanced image.py with new nodes like LoadImageFromPath and SaveImageToPath.
Updated server_endpoints.py with file-backed dictionary for user data persistence.
Refined node creation and sampling logic in pytorch_nodes.py and hf_diffusers_nodes.py for improved modularity and consistency."
Refined NODE_CLASS_MAPPINGS to include new classes like SelectStrFromDict, Text_ETK, and StringsToList.
Enhanced functionality for string and list operations in basic.py.
Introduced FuncAnysToDict in functional.py for dynamic dictionary generation from multiple inputs.
Removed redundant redefinitions of mappings in audio.py."
Key updates:
Added AddFramesToVideoFolder node to add frames to a video folder
Introduced VideoFolderProvider and VideoDefinitionProvider nodes to manage video folders and video definitions
Added FadeToInVideoFolder node for fading between frames in a video folder
Implemented ListedSVDConditioning node for applying SVD conditioning to a list of parameters
Added VideoFolderFFMPEGVideoInterpolate node to interpolate frames in a video folder using FFmpeg
Introduced AudioFolderProvider and AudioDefinitionProvider nodes for managing audio folders and definitions
Added AddAudioToAudioFolder and AddSilenceToAudioFolder nodes for adding audio files and silence to audio folders
Implemented GetAudioFilesLengthsFromAudioFolder node to retrieve lengths of audio files in a folder
Added AudioFolderRenderAudio node to join audio files in an audio folder
Introduced GenerateMusic node for generating music from descriptions using the AudioCraft library
Added XttsNode for text-to-speech conversion using the XTTS library
Implemented AdjustAudioVolumeForFiles node to adjust volume for a list of audio files
The changes also include the addition of LiteLLM nodes for text generation:
Added FuncAnysToText node that runs Python code with multiple input types and outputs a string
Introduced FuncAnysToLatent node for running code that outputs a latent representation
Added FuncAnysToString node for executing code with various inputs and producing a string output
Implemented FuncAnysToList node to run code that takes different input types and returns a list
Added FuncAnysToCond node for running code with multiple inputs and generating a conditioning output
These new nodes expand the capabilities for video and audio processing, allowing more advanced manipulations and generation of video and audio content within the graph.
This commit introduces a series of refactoring and enhancements across multiple modules, improving code readability, consistency, and adding new functionalities.
Highlights:
1. `tests/test_hf_diffusers_nodes.py`:
- Removed unnecessary whitespace.
- Improved variable assignment consistency (e.g., `ret=ret[0]` to `ret = ret[0]`).
- Added a new test case `test_create_unet2d_model` for Unet2d model creation.
2. `server_endpoints.py`:
- Cleaned up commented-out code and added more descriptive comments.
- Refactored the construction of `real_git_path` for better readability and future maintenance.
- Implemented minor formatting improvements for better code structure.
3. `basic.py`:
- Introduced a new class `TupleNode` to handle tuple data types.
- Modified `StrBreakout` to include dynamic typing for list conversion.
- Enhanced code readability with better variable naming and added comments for clarity.
4. `pytorch_nodes.py`:
- Added import statements within functions for `torch`.
- New classes `SaveTorchTensor` and `LoadTorchTensor` added for saving and loading torch tensors.
- `GridSearchTraining` class updated with minor formatting changes.
5. `image.py`:
- In `FloodFillNode`, removed a redundant category assignment.
- Added two new classes: `DecorateHeightMap` and `RandomImageFromFolder`. These classes provide functionalities for decorating height maps with images and selecting random images from a folder, respectively.
6. `hf_diffusers_nodes.py`:
- Fixed formatting in tensor format options and added a new line for clarity.
- New class `CreateUnet2dModel` added for creating a UNet2D model with customizable parameters.
- Introduced `TrainUnet2dModel` for training a UNet2D model with a dataset, including detailed setup for the training process.
7. `config.yaml`:
- Added a new base path for `gitio` to centralize the configuration.
These changes aim to improve the maintainability of the codebase, streamline testing processes, and expand the functionalities within the modules, particularly related to image processing and model training in PyTorch and diffusers.
This commit introduces significant enhancements and new features across multiple files in the project. Major changes include:
1. `pytorch_nodes.py`: Expanded PyTorch node functionalities with additional imports, device compatibility checks, optimizer state updates, and training enhancements. Introduced new methods like `optimizer_to` for transferring optimizer state to a specified device, `reset_optimizer_weights` to clear optimizer states, and `make_image_tensor` for converting outputs to image tensors. Also, improved the training process with device compatibility checks, optimizer updates, and the ability to save and load models during training for efficiency.
2. `server_endpoints.py`: Modified server endpoints for improved git integration and user management. Adjustments include dynamic handling of different git repositories, refined branch and user management in GitIO class, and the integration of user-specific git settings in user management functions.
3. `config.yaml`: Added new configurations for temporary directory and git repositories, providing more flexibility and control over file management and version control integration.
4. `torchvision_nodes.py`: Introduced a new file to integrate torchvision functionalities. It includes classes for transform composition, application of torchvision transforms, and handling of torchvision-specific data structures.
5. `hf_diffusers_nodes.py`: Integrated Hugging Face's diffusers library with new nodes for loading and utilizing DDPMPipelines, handling tensor to image conversions, and managing datasets from the Hugging Face ecosystem.
6. `image.py`, `video.py`, `functional.py`: Refactored and enhanced existing functionalities. Included minor tweaks and bug fixes to improve the overall efficiency and robustness of image and video processing capabilities.
7. `requirements.txt`: Updated dependencies, replacing `fastnoisesimd` with `diffusers` to align with the newly integrated Hugging Face diffusers functionalities.
8. `config.py`: Improved the configuration file loading mechanism with dynamic path handling and the ability to replace placeholders with actual paths, enhancing the flexibility of configuration management.
This comprehensive update significantly enhances the project's capabilities in machine learning, especially in PyTorch and torchvision, and improves server endpoint functionalities for better user and version control management.
Imports Repo from the git module.
Removes unused imports such as InvalidGitRepositoryError, os, copy, re, and imports from custom modules.
Removes the check_for_js_extension() function and its related logic, which checked for and copied JavaScript extension files.
image.py:
Modifies the ImageStackAndMatch class:
Updates the input type specifications.
Adds a new function hex_to_rgb to convert hex color codes to RGB tuples.
Adds a new function pad_to_max_size to pad images to a specified size.
Adds a new function resize_image to resize images with interpolation.
Modifies the process_image function to use the new functions for scaling and padding.
Modifies the process_images function to process up to four images, scaling them, and stacking them into a tensor.
functional.py:
Adds a new class FuncAnysToLatent:
This class runs eval on given text with five "any" inputs and outputs a latent value.
Specifies input types for five "any" inputs and the text to evaluate.
Provides a func function that executes the provided code with the inputs and returns the result as a latent value.
basic.py:
Modifies the get_value function:
Adds a check to ensure that LIST is a list type before returning it; otherwise, it returns an error message.
This file introduces a module containing various PyTorch-related classes and functionalities for use within the Eternal Toolkit (ETK) environment.
It imports essential PyTorch modules such as torch, torch.nn, and others.
Defines two dictionaries, NODE_CLASS_MAPPINGS and NODE_DISPLAY_NAME_MAPPINGS, for mapping class names to display names.
Implements a decorator ETK_pytorch_base to customize attributes of PyTorch-related classes.
Contains multiple PyTorch-related classes with the following key functionalities:
PyTorchDatasetDownloader: Downloads PyTorch datasets and returns them as tensors.
DatsetSplitter: Splits a dataset into training and validation sets.
AddLinearLayerNode: Adds linear layers to PyTorch models.
AddSoftmaxLayerNode: Adds softmax layers to PyTorch models.
SequentialModelProvider: Provides default sequential PyTorch models.
PyTorchInferenceNode: Performs inference using pre-trained PyTorch models.
FlattenTensor and FlattenDataset: Flatten tensors and datasets.
ReshapeTensor: Reshapes tensors.
ChangeTensorType: Changes the data type of tensors.
TensorToList: Converts tensors to lists.
SliceTensor: Slices tensors.
SaveModel and LoadModel: Save and load PyTorch models.
TrainModel: Trains PyTorch models on datasets.
TensorsToDataset and DatasetToDataloader: Convert tensors to datasets and datasets to data loaders, respectively.
Each class has well-defined input types, return types, and functions tailored to specific PyTorch operations.
Exception handling and error messages are provided to ensure robustness and usability.
The classes are designed to work seamlessly with both CPU and GPU devices based on the specified device parameter.
These classes extend the capabilities of ETK by providing powerful PyTorch functionalities for model development and data processing.
These new files and their contents enhance the functionality of ETK by enabling users to work with PyTorch models, datasets, and tensors more effectively. They offer a wide range of operations for deep learning and machine learning tasks within the ETK environment.
This file introduces a Python script for managing JavaScript extension files within a specific directory structure.
It includes an import statement for the os module.
Imports essential variables ETK_PATH and GDE_PATH from the custom_nodes.EternalKernelLiteGraphNodes.local_shared module.
Defines a function called check_for_js_extension(), which:
Verifies the existence of JavaScript extension files.
Compares modification timestamps to determine which file is newer.
Copies files between directories if they are outdated.
Provides user feedback during the copying process.
Calls the check_for_js_extension() function at the end of the script to perform the checks and copying.
Tasha Upchurch