#benchmark
2 related articles
·2 min
BatchBench: Evaluating Batch Effect and Domain Shift Robustness in Biomedical Images
BatchBench is a new leaderboard based on the JUMP-CP dataset that evaluates AI model robustness to batch effects and domain shift in biomedical images, advancing domain adaptation research.
Read more →
·3 min
Local Speech-to-Text on Mac: Whisper vs Parakeet Benchmark Comparison
The author of open-source dictation app Plainsay benchmarked Whisper large-v3-turbo vs Parakeet TDT 0.6B v3 on Apple Silicon: Parakeet averages 0.10 s latency and 2.50% WER, Whisper 0.68 s and 2.89% — the accuracy gap falls within normalization noise.
Read more →