hstdra / pandora-ai

Link to page

Mouth train data

Introduction

Click start button to generate data. You can talking or not talking, after that the data will be automatically capture. You can monitor last status (NORMAL/TALKING) and total number by category of generated data. At the end you can save the data by click save button.

Best practices:

You should use one emotion at on time for all the data.

Divide train data by do the test in multiple times and save JSON file like:

• BinhThuong_NORMAL.json
• MiengMoNhungKhongNoi_NORMAL.json
• Ho_NORMAL.json
• NoiTo_TALKING.json
• NoiNho_TALKING.json