names-whisper-en-spectrogram-new-method

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: 0.0374
• Ner percent: 98.7838
• Wer: 0.8407

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: 16
• 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	Ner percent	Wer
0.0039	5.0505	1000	0.0352	97.8378	1.1843
0.0005	10.1010	2000	0.0350	98.9189	0.8674
0.0003	15.1515	3000	0.0361	98.7838	0.8340
0.0002	20.2020	4000	0.0370	98.7838	0.8373
0.0002	25.2525	5000	0.0374	98.7838	0.8407

Framework versions

• Transformers 4.40.1
• Pytorch 2.2.2+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1