lhbing-ai

Error encountered during pytorch2onnx: AttributeError:'SegDecNet' Object has no attribute 'glob_max_lr_mlutiplier_mask'

16k model training configuration and 16k pre-trained model

1

Hi, I need to train a 16 kHz model for noise removal in the meeting scenario, and there are two questions: 1. If I train based on 16K audio data,...

端侧设备部署，onnx推理耗时20ms，是否正常以及还是否有优化手段

17

cpu cotext-a7双核@1.2G 在端侧部署了实时流，单帧（16ms数据）推理耗时约20ms，也就是在这款设备上达不到实时流 1.当前设备20ms的推理时间是否正常 2.模型上还能否进一步裁剪的可能，业务上有没有优化的手段，比如输入的数据大一点，现在输入的是256的数据，也就是推理要小于16ms才有落地可能

May I ask if it is necessary to add data on mute and noise synthesis for training purposes, and if there are any benefits tks

May I ask if it is necessary to add data on mute and noise synthesis for training purposes, and if there are any benefits