whisper-tiny-dv

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: 1.4917
• Wer Ortho: 58.5441
• Wer: 57.4380

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
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant_with_warmup
• lr_scheduler_warmup_steps: 50
• training_steps: 5000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
0.0013	17.86	500	1.0739	56.0148	55.3719
0.0003	35.71	1000	1.1575	54.3492	53.6009
0.0002	53.57	1500	1.2226	55.3979	54.7226
0.0001	71.43	2000	1.2711	56.6934	55.4900
0.0001	89.29	2500	1.3089	56.1999	55.1948
0.0	107.14	3000	1.3487	55.4596	54.4864
0.0	125.0	3500	1.3865	56.4466	55.5490
0.0	142.86	4000	1.4259	58.9759	57.6741
0.0	160.71	4500	1.4563	58.2973	57.0838
0.0	178.57	5000	1.4917	58.5441	57.4380

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1+cu117
• Datasets 2.13.1
• Tokenizers 0.13.3