base-asr-with-tpbs-cv1

This model is a fine-tuned version of Botnoi/base-asr-with-lm on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6284
• Wer: 0.2654
• Cer: 0.0774
• Clean Cer: 0.0712

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: 0.0001
• train_batch_size: 32
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer	Clean Cer
0.125	26.67	400	0.5910	0.2732	0.0841	0.0793
0.0426	53.33	800	0.6067	0.2647	0.0784	0.0725
0.0277	80.0	1200	0.6284	0.2654	0.0774	0.0712

Framework versions

• Transformers 4.27.4
• Pytorch 1.13.1+cu117
• Datasets 2.11.0
• Tokenizers 0.13.2