train03

This model is a fine-tuned version of openai/whisper-base on the ksponspeech_03 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4521
• Cer: 14.7492

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: 128
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 1
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
0.7999	0.13	100	0.6914	20.0296
0.5632	0.26	200	0.5602	16.9970
0.5309	0.39	300	0.5195	15.8134
0.4961	0.52	400	0.4940	15.6669
0.4865	0.65	500	0.4764	15.1768
0.4688	0.77	600	0.4607	14.9871
0.4357	0.9	700	0.4521	14.7492

Framework versions

• Transformers 4.36.0.dev0
• Pytorch 1.13.1+cu117
• Datasets 2.15.0
• Tokenizers 0.15.0