wangpuupup/test

Whisper Small dysarthric Dutch

This model is a fine-tuned version of qmeeus/whisper-small-nl on the data/copas copas-full dataset. It achieves the following results on the evaluation set:

• Loss: 0.3314
• Wer: 25.0816

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_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: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.3462	0.1	500	0.3721	29.9746
0.1205	1.1	1000	0.3286	26.4226
0.0642	2.1	1500	0.3139	25.2628
0.0443	3.1	2000	0.3169	25.0453
0.0356	4.1	2500	0.3226	25.0634
0.0321	5.09	3000	0.3237	24.7191
0.0317	6.09	3500	0.3322	24.7735
0.0347	7.09	4000	0.3355	25.0634
0.0357	8.09	4500	0.3364	25.2265
0.0318	9.09	5000	0.3314	25.0816

Framework versions

• Transformers 4.26.0.dev0
• Pytorch 1.12.1+cu116
• Datasets 2.4.0
• Tokenizers 0.12.1