whisper-tiny-aug-14-03-25

This model is a fine-tuned version of PhanithLIM/whisper-tiny-aug-29dec on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0688
• Wer: 70.9939

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• lr_scheduler_warmup_steps: 1000
• num_epochs: 9

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.1402	1.0	364	0.1076	82.2212
0.1014	2.0	728	0.0910	78.4741
0.0851	3.0	1092	0.0832	75.9403
0.0739	4.0	1456	0.0787	74.3304
0.0663	5.0	1820	0.0749	72.8278
0.0602	6.0	2184	0.0725	72.7998
0.0556	7.0	2548	0.0713	71.8852
0.051	8.0	2912	0.0697	71.8059
0.0476	9.0	3276	0.0688	70.9939

Framework versions

• Transformers 4.49.0
• Pytorch 2.4.0
• Datasets 3.4.0
• Tokenizers 0.21.1