jazzhong1/jazzhong1_medical_whisper_cut_small_3

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

• Loss: 0.2959
• Cer: 8.2065

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: 4e-05
• train_batch_size: 32
• 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: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
0.3011	0.9381	1000	0.3146	11.6877
0.2058	1.8762	2000	0.2831	9.7311
0.1046	2.8143	3000	0.2834	8.9407
0.0364	3.7523	4000	0.2867	8.5658
0.0111	4.6904	5000	0.2959	8.2065

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1