steja/whisper-small-yoruba

Whisper_small_Yoruba

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

• Loss: 1.6773
• Wer: 67.8866

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: 16
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• total_eval_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 2000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.013	36.35	400	1.4068	72.9681
0.0008	72.7	800	1.5546	68.4507
0.0003	109.09	1200	1.6400	67.9137
0.0002	145.43	1600	1.6773	67.8866
0.0002	181.78	2000	1.6901	68.1123

Framework versions

• Transformers 4.25.1
• Pytorch 1.13.0+cu117
• Datasets 2.7.1
• Tokenizers 0.13.2