whisper-tiny-finetuned-minds14

This model is a fine-tuned version of openai/whisper-tiny on the PolyAI/minds14 dataset. It achieves the following results on the evaluation set:

• Loss: 0.5352
• Wer Ortho: 0.3029
• Wer: 0.3046

Model description

I have made it for audio corse Unit 5 Hands-on. Here is some additional info https://outleys.site/en/development/AI/hugface-unit-5-excercise-guide/

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: 2e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• lr_scheduler_warmup_steps: 50
• num_epochs: 6

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
1.1444	0.8850	100	0.4740	0.3411	0.3388
0.2788	1.7699	200	0.4633	0.2986	0.3017
0.1377	2.6549	300	0.4969	0.3048	0.3052
0.0561	3.5398	400	0.5145	0.3017	0.3034
0.0177	4.4248	500	0.5241	0.3091	0.3117
0.01	5.3097	600	0.5352	0.3029	0.3046

Framework versions

• Transformers 4.46.2
• Pytorch 2.5.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3