whisper-large-v3-Cantonese-fine-tune-bible-1000

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: 0.4357
• Wer: 83.4483

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: 5e-05
• train_batch_size: 16
• 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: 100
• training_steps: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0673	7.6923	100	0.2826	99.3103
0.0276	15.3846	200	0.3737	82.0690
0.0174	23.0769	300	0.4343	89.6552
0.005	30.7692	400	0.4248	80.6897
0.0002	38.4615	500	0.4275	82.7586
0.0001	46.1538	600	0.4303	82.7586
0.0	53.8462	700	0.4326	83.4483
0.0	61.5385	800	0.4342	83.4483
0.0	69.2308	900	0.4352	83.4483
0.0	76.9231	1000	0.4357	83.4483

Framework versions

• Transformers 4.46.0.dev0
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.20.0