README.md

metadata

library_name: transformers
license: mit
base_model: openai/whisper-large-v3-turbo
tags:
  - generated_from_trainer
metrics:
  - wer
  - bleu
model-index:
  - name: whisper-large-v3-turbo-OpenSLR-GL-EN
    results: []
datasets:
  - juanjucm/OpenSLR-SpeechT-GL-EN
language:
  - gl

whisper-large-v3-turbo-OpenSLR-GL-EN

This model is a fine-tuned version of openai/whisper-large-v3-turbo on juanjucm/OpenSLR-SpeechT-GL-EN. It achieves the following results on the evaluation set:

• Loss: 0.8768
• Wer: 36.7059
• Bleu: 49.4165

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-06
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• total_train_batch_size: 32
• total_eval_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 8
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Bleu
0.5639	1.0	62	1.0983	41.4588	42.0181
0.2002	2.0	124	0.9388	41.6471	42.2327
0.0984	3.0	186	0.8720	37.5529	46.4185
0.0289	4.0	248	0.8828	38.6353	45.8710
0.028	5.0	310	0.8531	40.3294	44.8830
0.0101	6.0	372	0.8768	36.7059	49.4165
0.0085	7.0	434	0.9114	38.6353	47.3759
0.0059	8.0	496	0.9230	38.1176	47.2525

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.20.0