whisper-medium-aeb_AT

This model is a fine-tuned version of Rziane/whisper-medium-aeb_TunCS on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.2080
• Wer: 63.7618

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: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	1.0	99	1.9925	106.6050
No log	2.0	198	1.2492	68.9387
No log	3.0	297	1.0616	66.4557
No log	4.0	396	1.0217	65.0925
No log	5.0	495	1.0509	67.5755
1.3526	6.0	594	1.0961	64.5245
1.3526	7.0	693	1.1130	65.1899
1.3526	8.0	792	1.1661	62.6582
1.3526	9.0	891	1.1678	62.0253
1.3526	10.0	990	1.2080	63.7618

Framework versions

• Transformers 4.45.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.20.0