TTS Dataset Viewer

Visualize, find errors, and edit your TTS datasets

Easily find and edit repeated words and words not in CMU dictionary.

Installation

• Install dependencies

pip install streamlit st-annotated-text pandas

• Install forked g2p repo:

pip install g2p_en@git+https://github.com/uberduck-ai/g2p

• Clone this repo:

git clone git@github.com:uberduck-ai/dataset_viewer.git

• Run the app

cd dataset_viewer
streamlit run app.py

Usage Instructions

• Run the app: streamlit run app.py
• Enter the dataset information in the sidebar on the left
• NOTE: Please make a copy of the filelist... edits with this app will overwite the provided filelist
• Select Visualize in the form and the dataset will populate the main screen

Editing

• Select the Edit checkbox
• Make the desired edits in the textbox that pops up
• Save the edits by pressing Submit
• Unselect the Edit checkbox to see the edits and hide the editing widgets