# ALFFAAmharic a dataset for Amharic ASR

A Dataset script to access the ALFFAmharic through the transformer tool. This dataset allows easy loading and processing of the ALFFAmharic 
training audio and corresponding transcriptions using the transformer's load_dataset() method.

`amha = load_dataset("Samuael/ALFFAAmharic", data_dir="/path/to/the/data/folder/")`

Training folder structure:

```
    folder- train
          - wav/
            - 1234.wav
            - 5678.wav
            -  ...
          - text.txt
       - test
          - wav/
            - 1234.wav
            - 5678.wav
            -  ...
          - text.txt
```

---
license: other
task_categories:
- automatic-speech-recognition
language:
- am
pretty_name: AlFFAmharic
size_categories:
- n<1K
---