Sign in to view source links and access this dataset
Description
German-language speech recognition model evaluation results on multiple benchmark datasets. The dataset, created by user flozi00, presents a table of Word Error Rate (WER) scores for models like 'primeline-whisper-large-v3-turbo-german' across test sets including Tuda-De, multilingual librispeech, and common_voice_19_0. It was last updated on January 6, 2026.
Use Cases
Benchmarking German ASR model performance based on the reported Word Error Rate scores.
Comparing model accuracy across different German speech datasets like Tuda-De and Common Voice.
Selecting a pre-trained model for a German speech application based on its evaluated performance on specific test sets.
Strengths
Provides direct numerical evaluation scores (e.g., 2.62, 6.37 WER) for model comparison.
Evaluates models on multiple distinct German speech datasets, suggesting a multi-faceted performance view.
Limitations
Description metadata is limited; actual data quality requires manual inspection after download.
Column-level documentation is absent; field semantics must be inferred after download.
Row count and full data scale are unknown, which may limit suitability assessment.
Provenance
Source
huggingface user flozi00
Collection Method
Likely aggregated from model evaluation runs on public benchmark datasets.
Freshness
Last updated 2026-01-06 14:47:17; freshness should be verified.
Geography
Germany (language focus)
License is unknown; terms of use must be verified before application.