em_wav

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

• Loss: 1.6818
• Wer: 96.8954

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: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
3.9478	0.08	100	3.4707	99.1013
2.5514	0.16	200	2.4346	96.7116
1.7725	0.24	300	1.7743	99.9387
1.752	0.32	400	1.7586	97.1201
1.7447	0.4	500	1.7461	98.1413
1.7118	0.48	600	1.7304	97.1201
1.6823	0.56	700	1.7147	97.1201
1.7535	0.65	800	1.6987	97.8145
1.6772	0.73	900	1.6895	97.7941
1.6552	0.81	1000	1.6818	96.8954

Framework versions

• Transformers 4.28.1
• Pytorch 2.0.1+cu117
• Datasets 2.13.0
• Tokenizers 0.13.3