vilon888
vilon888
Hi, 请教一下,想基于wenet训练一个最终模型文件大小控制在80MB以内,我看了一下examples/aishell/s0/conf/train_u2++_efficonformer_v1_stream.yaml配置训练出来的模型pt文件也有近200MB, 请问wenet中有哪个配置训练出来的模型文件大小可以控制在80MB以内,且支持中文流式ASR。 官方[EfficientConformer small](https://github.com/burchim/EfficientConformer/blob/master/configs/EfficientConformerCTCSmall.json)配置的训练出来的模型大小只有[50MB](https://drive.google.com/drive/folders/1Dqu1RTHQ8jxGxEPar2-WMjR0hRhmkpoU)左右, 请问wenet 中的efficient conformer有支持这种mini模型的配置吗?或者需要从哪个conf配置修改,得到一个efficient conformer训练配置? 感谢!!!!
We appreciate you go through Apollo documentations and search previous issues before creating an new one. If neither of the sources helped you with your issues, please report the issue...