README.md

metadata

library_name: transformers
language:
  - sn
license: apache-2.0
base_model: openai/whisper-small
tags:
  - generated_from_trainer
datasets:
  - DigitalUmuganda/Afrivoice
metrics:
  - wer
model-index:
  - name: Whisper Small Shona - Beijuka Bruno
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Afrivoice_shona
          type: DigitalUmuganda/Afrivoice
        metrics:
          - name: Wer
            type: wer
            value: 0.2574299634591961

Whisper Small Shona - Beijuka Bruno

This model is a fine-tuned version of openai/whisper-small on the Afrivoice_shona dataset. It achieves the following results on the evaluation set:

• Loss: 0.5995
• Wer: 0.2574
• Cer: 0.0468

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e-05
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.5547	1.0	2204	0.3322	0.2800	0.0544
0.2437	2.0	4408	0.3013	0.2732	0.0572
0.1557	3.0	6612	0.3104	0.2698	0.0526
0.0954	4.0	8816	0.3244	0.2713	0.0520
0.0536	5.0	11020	0.3554	0.2719	0.0509
0.0297	6.0	13224	0.3843	0.2679	0.0505
0.0185	7.0	15428	0.4178	0.2679	0.0502
0.0132	8.0	17632	0.4407	0.2642	0.0502
0.0108	9.0	19836	0.4662	0.2653	0.0498
0.0087	10.0	22040	0.4724	0.2628	0.0498
0.0077	11.0	24244	0.4939	0.2652	0.0495
0.0068	12.0	26448	0.4969	0.2666	0.0503
0.0062	13.0	28652	0.5177	0.2681	0.0495
0.006	14.0	30856	0.5313	0.2694	0.0513
0.0053	15.0	33060	0.5357	0.2644	0.0491
0.0049	16.0	35264	0.5420	0.2636	0.0499
0.0042	17.0	37468	0.5464	0.2620	0.0503
0.0042	18.0	39672	0.5526	0.2563	0.0485
0.004	19.0	41876	0.5666	0.2639	0.0501
0.0038	20.0	44080	0.5678	0.2631	0.0501
0.0036	21.0	46284	0.5750	0.2672	0.0495
0.0036	22.0	48488	0.5739	0.2562	0.0490
0.0032	23.0	50692	0.5869	0.2639	0.0499
0.003	24.0	52896	0.5912	0.2625	0.0491
0.003	25.0	55100	0.5896	0.2596	0.0494
0.0032	26.0	57304	0.5945	0.2596	0.0495
0.0025	27.0	59508	0.6071	0.2630	0.0525
0.0025	28.0	61712	0.6165	0.2560	0.0483

Framework versions

• Transformers 4.45.2
• Pytorch 2.2.0+cu121
• Datasets 3.0.1
• Tokenizers 0.20.1