Mal_ASR_Whisper_small_imasc_1000

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: 0.0642
• Wer: 52.2853

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: 16
• 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: 2000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.3098	0.74	200	0.2613	200.6810
0.1009	1.48	400	0.0988	54.5952
0.0559	2.22	600	0.0722	44.6184
0.0518	2.96	800	0.0608	39.1631
0.0285	3.7	1000	0.0573	46.0858
0.0166	4.44	1200	0.0567	46.7036
0.0082	5.19	1400	0.0589	50.9513
0.0075	5.93	1600	0.0590	65.6252
0.0031	6.67	1800	0.0629	57.2913
0.0018	7.41	2000	0.0642	52.2853

Framework versions

• Transformers 4.36.0.dev0
• Pytorch 2.0.1+cu117
• Datasets 2.12.0
• Tokenizers 0.14.0