afrispeech_ayarma_large-v2 – AI Model by Ajegetina | AlphaNeural AI

You can deploy this model and start earning money today!

Ajegetina/
afrispeech_ayarma_large-v2

like

0

automatic-speech-recognition

generated_from_trainer

tobiolatunji/afrispeech-200

openai/whisper-large-v2

endpoints_compatible

Views

No views yet

Whisper Large v2 - Ayarma

This model is a fine-tuned version of openai/whisper-large-v2 on the Afrispeech 200 dataset. It achieves the following results on the evaluation set:

Loss: 0.2719
Wer: 11.7674

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: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 16
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 500
training_steps: 4000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.4933	0.125	500	0.4292	15.6823
0.3863	0.25	1000	0.3523	13.9511
0.4259	0.375	1500	0.3369	13.3288
0.3869	0.5	2000	0.3361	13.5438
0.4432	0.625	2500	0.2995	12.4010
0.3421	0.75	3000	0.2926	12.2539
0.2882	0.875	3500	0.2738	11.6316
0.2725	1.1125	4000	0.2719	11.7674

Framework versions

Transformers 4.52.0
Pytorch 2.9.1+cu128
Datasets 2.19.0
Tokenizers 0.21.4