whisper-small-ft-balbus-sep28k-multiclass_v3a-ft-balbus-sep28k-multiclass_v4_IT

This model is a fine-tuned version of miosipof/whisper-small-ft-balbus-sep28k-multiclass_v3a on the Apple dataset dataset. It achieves the following results on the evaluation set:

• Loss: 0.2301
• Accuracy: 0.8592
• Precision: 0.8795
• Recall: 0.8594
• F1: 0.8604
• Roc-auc: None

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: 3e-06
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.5
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Roc-auc
0.6519	0.4188	100	0.6127	0.4741	0.5574	0.4759	0.4098	None
0.4774	0.8377	200	0.4080	0.7468	0.7651	0.7470	0.7497	None
0.2425	1.2565	300	0.2518	0.8439	0.8566	0.8443	0.8445	None
0.2275	1.6754	400	0.2301	0.8592	0.8795	0.8594	0.8604	None

Framework versions

• Transformers 4.45.2
• Pytorch 2.2.0
• Datasets 3.6.0
• Tokenizers 0.20.3