kiranpantha/whisper-tiny-ne

Whisper Medium - Kiran Pantha

This model is a fine-tuned version of openai/whisper-tiny on the OpenSLR54 dataset. It achieves the following results on the evaluation set:

• Loss: 0.2728
• Wer: 53.8899

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: 8
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.6656	0.1200	500	0.6245	85.4211
0.4586	0.2399	1000	0.4490	73.6369
0.3772	0.3599	1500	0.3930	68.5879
0.3437	0.4798	2000	0.3498	63.7222
0.3214	0.5998	2500	0.3279	61.2297
0.3186	0.7198	3000	0.3095	59.3696
0.2965	0.8397	3500	0.2930	56.9504
0.2759	0.9597	4000	0.2825	56.0249
0.2474	1.0797	4500	0.2758	54.5588
0.2195	1.1996	5000	0.2728	53.8899

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1