procit006/stt_nl_sept8

Whisper Small

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

• Loss: 0.0195
• Wer: 1.2366

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: 3e-05
• train_batch_size: 8
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 150
• training_steps: 3000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0894	0.2040	500	0.0849	5.4565
0.0562	0.4079	1000	0.0476	3.1760
0.0349	0.6119	1500	0.0327	2.2623
0.0318	0.8158	2000	0.0254	1.5854
0.0069	1.0198	2500	0.0208	1.3277
0.0059	1.2237	3000	0.0195	1.2366

Framework versions

• Transformers 4.42.3
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1