Open zsLin177 opened 5 days ago
Hello,
You can refer to this Contextualized Streaming End-to-End Speech Recognition with Trie-Based Deep Biasing and Shallow Fusion and github to calculate the U-WER, B-WER. The scripts for calculating Recall needs to be organized later. Thank you for your interest in our work.
Hello,
Thanks for releasing the dataset. I am conducting some experiments based on your dataset and I am curious about the calculations for U-WER, B-WER, and Recall. Could you please provide the script for these calculations? 🙏