VoiceMath-Tiny-Quad

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

• Loss: 0.4370
• Wer: 11.5062

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: 8
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	1.0	100	0.3667	17.2346
No log	2.0	200	0.3248	11.3086
No log	3.0	300	0.3718	12.4444
No log	4.0	400	0.3725	12.0
0.1807	5.0	500	0.3931	11.6049
0.1807	6.0	600	0.4150	11.7531
0.1807	7.0	700	0.4069	11.5556
0.1807	8.0	800	0.4145	11.5062
0.1807	9.0	900	0.4177	11.6543
0.0012	10.0	1000	0.4221	11.6049
0.0012	11.0	1100	0.4247	11.5062
0.0012	12.0	1200	0.4270	11.5062
0.0012	13.0	1300	0.4293	11.6049
0.0012	14.0	1400	0.4314	11.6543
0.0002	15.0	1500	0.4329	11.6543
0.0002	16.0	1600	0.4342	11.6543
0.0002	17.0	1700	0.4354	11.5062
0.0002	18.0	1800	0.4363	11.5062
0.0002	19.0	1900	0.4368	11.5062
0.0002	20.0	2000	0.4370	11.5062

Framework versions

• Transformers 4.30.0.dev0
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.13.3