pratyush_whisper_tiny_finetune_teacher_babble_noise_libri360_epochs_50_batch_8_try2

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

• Loss: 0.3760
• Wer: 16.4737

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 512
• total_train_batch_size: 4096
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.2
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.872	5.98	150	0.5003	22.0157
0.2818	11.98	300	0.3427	17.1317
0.1575	17.98	450	0.3189	16.1630
0.1037	23.98	600	0.3261	16.0472
0.0732	29.98	750	0.3416	16.1354
0.0547	35.98	900	0.3575	16.2237
0.0439	41.98	1050	0.3695	16.3854
0.0381	47.98	1200	0.3760	16.4737

Framework versions

• Transformers 4.24.0
• Pytorch 1.12.1
• Datasets 2.7.0
• Tokenizers 0.11.0