no-voice-clone-large-finetune

This model is a fine-tuned version of openai/whisper-large-v3 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4678
• Wer: 18.7667

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• training_steps: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0528	1.8692	100	0.4677	20.6937
0.0076	3.7383	200	0.4470	18.0848
0.0012	5.6075	300	0.4580	18.0255
0.0002	7.4766	400	0.4565	17.4326
0.0001	9.3458	500	0.4601	18.7370
0.0001	11.2150	600	0.4634	18.5295
0.0	13.0841	700	0.4653	18.5888
0.0	14.9533	800	0.4667	18.5591
0.0	16.8224	900	0.4675	18.7963
0.0	18.6916	1000	0.4678	18.7667

Framework versions

• Transformers 4.45.2
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.20.3