Enhance run_training function in app.py to ensure child Python processes are unbuffered for real-time logging. Update _prepare_script to include unbuffered flag and modify UI launch to support ssr_mode, improving compatibility with different Gradio versions.
Refactor UI in app.py to implement a tabbed layout for the Trainer and Prompt Generator sections, enhancing organization and user experience. The update includes restructured input fields and improved accessibility for image uploads and prompt generation functionalities.
Enhance _list_checkpoints function in app.py to skip files that are zero-length, too new, or unreadable. This improvement ensures more robust file handling by filtering out potentially in-flight files, thereby reducing errors during checkpoint listing.
Add zero-length file check in _list_checkpoints function of app.py to skip empty files during checkpoint listing. This enhancement improves the robustness of file handling by preventing errors related to empty files.
Refactor error handling in run_training function of app.py to remove unnecessary None yields. This change streamlines the output by ensuring consistent return values when errors occur, enhancing clarity in the training process.
Update prefix/suffix naming in _prepare_script of app.py to ensure consistency with target parameters. This change improves clarity in metadata creation by correctly labeling the flags for target prefix and suffix.
Refactor UI layout in app.py by reorganizing the prefix/suffix input fields into accordions for improved accessibility and visual coherence. This update enhances the user experience during image uploads by maintaining a consistent structure across control sections.
Refactor UI layout in app.py to improve organization of target image input fields. Adjust the arrangement of the prefix and suffix textboxes for better accessibility and visual coherence, enhancing the overall user experience during image uploads.
Refactor UI in app.py to enhance layout and styling of sections, improving organization and accessibility. Adjust padding and margins for better visual consistency, and update accordion titles for clarity on prefix/suffix functionality in image uploads.
Refactor UI layout in app.py by consolidating input fields into accordions for better organization and accessibility. This update enhances the user experience for managing target and control images, improving the overall interface for image uploads and training configurations.
Refactor UI layout in app.py to improve organization and accessibility of input fields by grouping related elements and adding custom CSS styling. This enhances the overall user experience for image uploads and training parameter configurations.
Refactor UI layout in app.py to enhance user experience by incorporating an accordion for control_0 image previews. This change improves organization and accessibility of image upload and preview functionalities.
Add hyperparameter overriding functionality in _prepare_script of app.py to allow safer and more flexible adjustments of learning rate, network dimension, seed, max training epochs, and save frequency. This enhances configurability for training scripts.
Refactor app.py to update prefix/suffix naming conventions for metadata creation and enhance UI with new training hyperparameter inputs. Modify train_QIE.sh to utilize dynamic hyperparameter values for training execution, improving configurability and user experience.
Add hyperparameter override options in _prepare_script function of app.py to allow dynamic adjustment of learning rate, network dimension, and seed during script preparation. Update UI elements for clarity in target image handling.
Implement target prefix/suffix functionality in app.py and create_image_caption_json.py for improved image processing. Update UI to reflect changes, enhancing user experience with clearer labeling and file handling. Refactor related functions to accommodate new naming conventions.
Add checkpoint listing functionality in app.py to track and display model checkpoints during training. Update run_training to yield checkpoint information and enhance Gradio UI with checkpoint file outputs for improved user experience.
Refactor app.py to set output_dir_base to the unique dataset directory and enhance Gradio UI launch by allowing access to specific runtime directories, improving file serving capabilities.
Update app.py to remove outdated startup instructions from the UI, enhance train_QIE.sh to dynamically set cache_directory, and modify requirements.txt to specify a custom index URL for torchvision installation.
Refactor logging in run_training function of app.py to accumulate log messages in a buffer before yielding, improving output consistency and readability during training execution.
Add dependency installation for musubi-tuner in app.py during startup. Implement best-effort upgrade for pip, setuptools, and wheel, and handle optional Torch extras for editable installation. Fallback to plain install if editable fails, ensuring smoother setup process.
Enhance image upload handling in app.py by implementing PNG conversion during file saving and adding prefix/suffix options for metadata creation. Update UI to include fields for main and control image prefixes/suffixes, improving user experience and flexibility in image processing.
Refactor UI in app.py to clearly separate mandatory control_0 image uploads from optional control images. Control_0 uploads are now displayed outside of the accordion for improved user guidance, while optional controls are organized within a closed accordion by default.
Update run_training function in app.py to enforce mandatory control_0 image uploads, improving error handling and user guidance. Modify UI to reflect that control_0 images are required, enhancing clarity for users during image uploads.
Refactor run_training function in app.py to return tuples for error and log messages, enhancing clarity in output handling. Update UI to remove user input for models_root, output_dir_base, and dataset_config, which are now resolved at runtime. Implement logic to locate the latest LoRA file for download after training completion.
Enhance app.py by adding an override option for run name in the _prepare_script function and updating the run_training function to use an auto-generated dataset directory name. Modify UI to reflect changes from dataset_name to output_name for better clarity in user input.
Refactor run_training function in app.py to simplify parameters and enhance control image handling. Update UI for image uploads and ensure at least one control folder is provided for training. Adjust data root resolution to auto-decide based on user input.
Reorder startup functions in app.py to ensure musubi-tuner is cloned before downloading models. This change maintains the intended execution flow during application startup.
Enhance app.py with new functions for handling data uploads and directory management. Implement data root resolution and control set preparation for training. Update main execution flow to accommodate new upload handling and directory structure.
Refactor app.py to support runtime-resolved paths for musubi-tuner and models directories. Implement fallback mechanisms for directory creation and cloning operations. Update UI components to reflect new runtime paths.
Update default paths in app.py and download_qwen_image_models.py to use /data and /auto directories. Adjust metadata generation script to reflect new working directory structure.
Enhance training functionality with ZeroGPU support and UI adjustments. Added options to override max epochs and save frequency, and implemented GPU request handling for Spaces compatibility.
