whisper-large-v3-genbed-all

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

• Loss: 0.4136
• Wer: 29.1573

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: 1.75e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 2500
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.614	0.3300	250	0.6284	55.5217
0.5843	0.6601	500	0.5559	45.7546
0.5024	0.9901	750	0.4794	39.4905
0.2772	1.3201	1000	0.4712	39.8964
0.3159	1.6502	1250	0.4251	37.8511
0.2998	1.9802	1500	0.4008	32.9488
0.1497	2.3102	1750	0.4105	31.5123
0.1412	2.6403	2000	0.3944	31.5551
0.1325	2.9703	2250	0.3839	30.5084
0.045	3.3003	2500	0.4136	29.1573

Framework versions

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