feat: Add a JSON output tab to display raw inference results for both Document and Visual Language tasks.

feat: Add early exit conditions to math escaping, prevent duplicate visualizations, and introduce initial loading indicators for inference.

feat: Integrate Gradio progress bar support and refine error handling in inference functions.

progres report fix

feat: Implement streaming OCR results by converting `run_inference` to a generator and updating UI wrappers to consume partial outputs.

feat: Add default configuration for the PaddleOCR-VL-1.5 pipeline.

refactor: Update PaddleOCR-VL model to 1.5 in configuration generation and remove the default.yaml file.

feat: Add Gradio progress bar updates to the `run_inference` function for better user feedback.

refactor: Update pipeline predict call to use keyword argument and iterate through results for printing.

refactor: Migrate image inputs from `gr.File` to `gr.Image` for native preview support, removing custom preview logic, and add full progress display for inference tasks.

feat: add logging for prediction pipeline output type and content

config: Update OCR model_name from TachiwinOCR-VL-1.5 to PaddleOCR-VL-1.5-0.9B.

chore: Update OCR_VL_Recognition model name from PaddleOCR-VL-0.9B to TachiwinOCR-VL-1.5.

feat: Download OCR model locally from Hugging Face, update pipeline configuration to use the local path, and add `archive` to `.gitignore`.

feat: Add Gradio UI and inference logic for audio language classification with a custom Whisper encoder.

refactor: Streamline PaddleOCR-VL pipeline setup using `paddlex` CLI for config generation and add new UI reference and existing OCR app files.

refactor: Improve PaddleOCR pipeline setup, configuration, and error handling in `app.py` and add new UI-related reference modules.

first commit

initial commit