This dataset was created by Alvenir in order to evaluate ASR models in Danish. It can also be used for training but the amount is very limited.
The dataset consists of .wav files with corresponding reference text. The amount of data is just above 5 hours spread across 50 speakers with age in the interval 20-60 years old. The data was collected by a third party vendor through their software and people. All recordings have… See the full description on the dataset page: https://huggingface.co/datasets/Alvenir/alvenir_asr_da_eval.
