whisper-medium-aeb_TunCS

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

• Loss: 0.5447
• Wer: 29.5377

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	163	0.6260	44.0925
No log	2.0	326	0.4777	42.5942
No log	3.0	489	0.4481	39.5548
0.6281	4.0	652	0.4537	39.7260
0.6281	5.0	815	0.4666	40.4538
0.6281	6.0	978	0.5000	33.3904
0.0853	7.0	1141	0.5104	29.7517
0.0853	8.0	1304	0.5232	30.6079
0.0853	9.0	1467	0.5309	30.0514
0.0114	10.0	1630	0.5447	29.5377

Framework versions

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