Wav2vec2 large xlsr cantonese This model was based on wav2vec2 large xlsr 53, finetuned using Common Voice/zh HK/6.1.0. The training code is similar to user ctl, except that the number of training epochs was 80 (doubled) and fp16 backend is apex. The model was trained using a single RTX 3090 and docker image is nvidia/cuda:11.1 cudnn8 devel. CER is 15.11% when evaluate against common voice zh HK test set. Result (CER) 15.11% Source Code See this GitHub Repo cantonese selfish project and demo video. Usage
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy