whisper-med-asd_v2

This model is a fine-tuned version of openai/whisper-medium on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5457
• Wer: 37.9455

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 1000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.1323	10.53	100	0.5968	71.875
0.0052	21.05	200	0.7108	77.7778
0.0005	31.58	300	0.7444	81.25
0.0004	42.11	400	0.7615	80.9028
0.0003	52.63	500	0.7780	79.8611
0.0003	63.16	600	0.7941	80.9028
0.0003	73.68	700	0.8077	80.9028
0.0003	84.21	800	0.8194	79.1667
0.0003	94.74	900	0.8276	79.5139
0.0003	105.26	1000	0.8305	79.1667

Framework versions

• Transformers 4.31.0.dev0
• Pytorch 2.0.1+cu117
• Datasets 2.13.1
• Tokenizers 0.13.3