README1.md

Audio Upload Structure (S2TT)

Follow this folder structure strictly while uploading audio files.

1. Base Path

S2TT/training_audio/

2. Language Folder

• Create a folder with the language name in lowercase only

Example:

maithili

3. Dataset Folder

Inside the language folder, create:

audio_indicvoices_r

This depends on the dataset you choose Example: audio_rasa, audio_indicvoices_r, etc.

4. Chunk Folder

Inside the dataset folder:

chunk_01

5. Audio Files

Inside chunk_01, place audio files:

audio_000000.wav
audio_000001.wav
...

Final Structure

S2TT/
└── training_audio/
    └── maithili/
        └── audio_indicvoices_r/
            └── chunk_01/
                ├── audio_000000.wav
                ├── audio_000001.wav
                └── ...

Important Rules

• Language name must be lowercase
• Folder name must be exactly: audio_indicvoices_r (depends on dataset, e.g., audio_rasa)
• Chunk name must be exactly: chunk_01
• Audio files must be in .wav format