whisper-large-v3-Punjabi-Version1

This model is a fine-tuned version of openai/whisper-large-v3 on the fleurs dataset. It achieves the following results on the evaluation set:

• Loss: 0.1883
• Wer: 44.8199

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: 3e-06
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 2000
• training_steps: 20000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.2523	7.4627	2000	0.3047	62.7791
0.1706	14.9254	4000	0.2324	52.0393
0.1466	22.3881	6000	0.2120	49.2781
0.1411	29.8507	8000	0.2019	47.2388
0.1294	37.3134	10000	0.1962	46.3456
0.1155	44.7761	12000	0.1926	45.5716
0.1196	52.2388	14000	0.1905	44.9539
0.1111	59.7015	16000	0.1889	44.8199
0.1066	67.1642	18000	0.1883	44.5743
0.1138	74.6269	20000	0.1883	44.8199

Framework versions

• PEFT 0.12.1.dev0
• Transformers 4.45.0.dev0
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1