whisper-large-v3-music-detection

This model is a fine-tuned version of openai/whisper-large-v3 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6421
• Accuracy: 0.7616
• F1: 0.7462
• Precision: 0.7670
• Recall: 0.7616
• F1 Pure Speech: 0.6735
• F1 Speech With Music: 0.5
• F1 Pure Music: 0.9697

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: 16
• eval_batch_size: 32
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall	F1 Pure Speech	F1 Speech With Music	F1 Pure Music
1.0164	1.3158	100	0.9843	0.4238	0.2523	0.1796	0.4238	0.0	0.0	0.5953
0.8459	2.6316	200	0.8073	0.4238	0.2523	0.1796	0.4238	0.0	0.0	0.5953
0.6761	3.9474	300	0.6421	0.7616	0.7462	0.7670	0.7616	0.6735	0.5	0.9697

Framework versions

• Transformers 4.57.1
• Pytorch 2.5.1+cu121
• Datasets 4.3.0
• Tokenizers 0.22.1