README.md

metadata

license: apache-2.0
tags:
  - afro-digits-speech
datasets:
  - crowd-speech-africa
metrics:
  - accuracy
model-index:
  - name: afrospeech-wav2vec-run
    results:
      - task:
          name: Audio Classification
          type: audio-classification
        dataset:
          name: Afro Speech
          type: chrisjay/crowd-speech-africa
          args: 'no'
        metrics:
          - name: Validation Accuracy
            type: accuracy
            value: 0.8

afrospeech-wav2vec-run

This model is a fine-tuned version of facebook/wav2vec2-base on the crowd-speech-africa, which was a crowd-sourced dataset collected using the afro-speech Space. It achieves the following results on the validation set:

• F1: 0.8
• Accuracy: 0.8

The confusion matrix below helps to give a better look at the model's performance across the digits. Through it, we can see the precision and recall of the model as well as other important insights.

Training and evaluation data

The model was trained on a mixed audio data from Rundi (run).

• Size of training set: 16
• Size of validation set: 5

Below is a distribution of the dataset (training and valdation)

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 3e-05
• train_batch_size: 64
• eval_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• num_epochs: 150

Training results

Training Loss	Epoch	Validation Accuracy
0.00183	1	0.6
0.0003991	50	0.8
0.0002174	100	0.6
0.0043911	150	0.4

Framework versions

• Transformers 4.21.3
• Pytorch 1.12.0
• Datasets 1.14.0
• Tokenizers 0.12.1