metadata

language:
  - de
license: apache-2.0
library_name: peft
tags:
  - generated_from_trainer
base_model: openai/whisper-small
datasets:
  - rmacek/ORF-whisper-small
metrics:
  - wer
model-index:
  - name: Whisper ORF Bundeslaender
    results:
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: ORF Bundesländer Voices
          type: rmacek/ORF-whisper-small
          args: 'config: de, split: test'
        metrics:
          - type: wer
            value: 26.328635352109302
            name: Wer

Whisper ORF Bundeslaender

This model is a fine-tuned version of openai/whisper-small on the ORF Bundesländer Voices dataset. It achieves the following results on the evaluation set:

Loss: 0.6154
Wer: 26.3286

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: 0.0001
train_batch_size: 16
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 500
training_steps: 4000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.5731	1.7153	1000	0.6095	36.1178
0.4196	3.4305	2000	0.6040	26.1875
0.4214	5.1458	3000	0.6086	26.3813
0.3579	6.8611	4000	0.6154	26.3286

Framework versions

PEFT 0.10.1.dev0
Transformers 4.40.0
Pytorch 2.2.1+cu121
Datasets 2.19.0
Tokenizers 0.19.1