# Speech Recognition on German video dataset (clean) Source: [Speech Recognition on German video dataset (clean)]() Collection index: [llms.txt]() Data updated: 2026-02-26T06:08:24.739Z Evaluation of Speech Recognition performance on the clean split of the German video dataset, typically measured by Word Error Rate (WER) and Word Error Rate Ratio (WERR). A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Speech Recognition | | Dataset (including any stated split) | German video dataset (clean) | | Metric | WER | | Unit | PERCENT | | Direction | Lower is better | | Results loaded | 9 of 9 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [WER](): Lower is better; unit: PERCENT. - [WERR](): Lower is better; unit: SCALAR. ## Recorded results | Method | WER | Result date | Report paper | | --- | --- | --- | --- | | + ce pretrain + ce loss, mid, top | 11.2 | Not recorded | 68fa68bc1e209131cea1de7f | | + aux loss | 11.3 | Not recorded | 68fa68bc1e209131cea1de7f | | + aux + kl loss | 11.3 | Not recorded | 68fa68bc1e209131cea1de7f | | + crosslingual pretrain + aux + kl loss | 11.3 | Not recorded | 68fa68bc1e209131cea1de7f | | + ce pretrain + ce loss, mid | 11.3 | Not recorded | 68fa68bc1e209131cea1de7f | | + crosslingual pretrain | 11.4 | Not recorded | 68fa68bc1e209131cea1de7f | | + kl loss | 11.5 | Not recorded | 68fa68bc1e209131cea1de7f | | + ce pretrain | 11.5 | Not recorded | 68fa68bc1e209131cea1de7f | | baseline | 11.6 | Not recorded | 68fa68bc1e209131cea1de7f | ### + ce pretrain + ce loss, mid, top WER: 11.2. Result ID: 698d81748b517eb43322062c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"lambda\_ce":0.6}. This record does not assert independent reproduction. ### + aux loss WER: 11.3. Result ID: 698d81748b517eb433220625. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"lambda\_aux":0.3}. This record does not assert independent reproduction. ### + aux + kl loss WER: 11.3. Result ID: 698d81748b517eb433220627. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"lambda\_aux":0.3}. This record does not assert independent reproduction. ### + crosslingual pretrain + aux + kl loss WER: 11.3. Result ID: 698d81748b517eb433220629. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"lambda\_aux":0.3}. This record does not assert independent reproduction. ### + ce pretrain + ce loss, mid WER: 11.3. Result ID: 698d81748b517eb43322062b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"lambda\_ce":0.6}. This record does not assert independent reproduction. ### + crosslingual pretrain WER: 11.4. Result ID: 698d81748b517eb433220628. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### + kl loss WER: 11.5. Result ID: 698d81748b517eb433220626. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"lambda\_aux":0.3}. This record does not assert independent reproduction. ### + ce pretrain WER: 11.5. Result ID: 698d81748b517eb43322062a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### baseline WER: 11.6. Result ID: 698d81748b517eb433220624. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.