whisper_tiny_char

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

• Loss: 0.3979
• Wer: 100.0

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: 0.0001
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant
• lr_scheduler_warmup_steps: 1000
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.1664	1.0	859	0.5140	100.0
0.5122	2.0	1718	0.4044	100.0
0.4035	3.0	2577	0.3596	100.0
0.3403	4.0	3436	0.3394	100.0
0.2967	5.0	4295	0.3395	100.0
0.2582	6.0	5154	0.3314	100.0
0.2286	7.0	6013	0.3571	100.0
0.2034	8.0	6872	0.3590	100.0
0.181	9.0	7731	0.3706	100.0
0.1618	10.0	8590	0.3979	100.0

Framework versions

• Transformers 4.44.0
• Pytorch 2.3.1
• Datasets 2.21.0
• Tokenizers 0.19.1