Makkoen/whisper-large-v3-cit-do015-wd0-lr1e-06-steps500-PHX

FULL6

This model is a fine-tuned version of openai/whisper-large-v3 on the 1003 FULL-2024-11-22 dataset. It achieves the following results on the evaluation set:

• Loss: 0.5210
• Wer Ortho: 25.9578
• Wer: 21.2796

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-06
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• gradient_accumulation_steps: 4
• 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: 300
• training_steps: 500
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
0.8252	3.5398	200	0.5752	28.6902	23.7035
0.4751	7.0796	400	0.5210	25.9578	21.2796

Framework versions

• Transformers 4.44.0
• Pytorch 1.13.1+cu117
• Datasets 2.21.0
• Tokenizers 0.19.1