ResearchDatasetsAverage across modelsAverage across modelsFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsMulti-path Speculative DecodingAverage across models (Qwen, Gemma, Llama) (test)16.69Throughput (tokens/s)6VLA HijackingAverage across Models Main Results (5 tasks)52.54ASR5
Multi-path Speculative DecodingAverage across models (Qwen, Gemma, Llama) (test)16.69Throughput (tokens/s)6