whisper-large-v3-Bengali-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: 0.1519
• Wer: 44.5003

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: 1000
• training_steps: 20000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.2412	4.8193	2000	0.2103	56.7927
0.208	9.6386	4000	0.1844	51.9640
0.195	14.4578	6000	0.1719	49.3167
0.194	19.2771	8000	0.1647	47.7358
0.1762	24.0964	10000	0.1597	46.6144
0.1763	28.9157	12000	0.1567	45.8361
0.168	33.7349	14000	0.1544	45.1192
0.1623	38.5542	16000	0.1530	44.8802
0.1601	43.3735	18000	0.1521	44.7822
0.1666	48.1928	20000	0.1519	44.5003

Framework versions

• PEFT 0.12.1.dev0
• Transformers 4.45.0.dev0
• Pytorch 2.4.1+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1