whisper-medium-cdsd1h

This model is a fine-tuned version of openai/whisper-medium on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.7572
• Wer: 65.6222
• Cer: 38.0119

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: 16
• 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: 100
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.7382	1.0	559	0.7516	72.6947	43.0148
0.3646	2.0	1118	0.7098	68.0394	43.5298
0.1402	3.0	1677	0.7165	66.1594	41.8050
0.0364	4.0	2236	0.7466	65.7565	38.2735
0.0102	5.0	2795	0.7572	65.6222	38.0119

Framework versions

• Transformers 4.44.2
• Pytorch 2.0.0+cu118
• Datasets 2.19.1
• Tokenizers 0.19.1