Created by: nemo9903

I remember the deepspeech worked before. But why it doesn't work now? As the training going on, the used memory increase along. At last, the system halt. I tried both pip and docker installation method. On top of ubuntu 14.04,16.04,18.04. And I also tried cpu /gpu versions paddle of 0.10,0.14,0.15,1.2.0,1.2.1. I also tried to modifiy the training parameters, for example, gru/layer size/batch size. I tried all of the configurations,but no use.

At last, the packages listed below. Is one of the configurations I tested.

Package Version

adium-theme-ubuntu 0.3.4 backports.functools-lru-cache 1.5 cffi 1.11.5 cycler 0.10.0 Cython 0.29.5 enum34 1.1.6 funcsigs 1.0.2 graphviz 0.10.1 kiwisolver 1.0.1 LinkChecker 9.3 llvmlite 0.27.0 matplotlib 2.2.3 nltk 3.4 numba 0.42.0 numpy 1.16.1 opencv-python 4.0.0.21 paddlepaddle 1.2.1 Pillow 5.4.1 pip 19.0.2 protobuf 3.1.0 pycparser 2.19 pyparsing 2.3.1 python-dateutil 2.8.0 python-speech-features 0.6 pytz 2018.9 rarfile 3.0 recordio 0.1.5 requests 2.9.2 resampy 0.2.1 scipy 1.2.1 setuptools 20.7.0 singledispatch 3.4.0.3 six 1.12.0 SoundFile 0.10.2 subprocess32 3.5.3 swig-decoders 1.1 unity-lens-photos 1.0 wheel 0.29.0