fine_tuned_asr_model

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

• Loss: 0.0906
• Cer: 5.7861

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
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 15.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
0.524	1.4706	100	0.0896	4.7486
0.1451	2.9412	200	0.0907	4.9049
0.0913	4.4118	300	0.0863	4.9947
0.0587	5.8824	400	0.0857	5.8460
0.0659	7.3529	500	0.1056	10.0725
0.0551	8.8235	600	0.0895	6.0555
0.0225	10.2941	700	0.0991	5.5899
0.0281	11.7647	800	0.0900	5.3006
0.022	13.2353	900	0.0887	6.4013
0.0213	14.7059	1000	0.0906	5.7861

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.19.1