alumae / gst-kaldi-nnet2-online

GStreamer plugin around Kaldi's online neural network decoder
Apache License 2.0
185 stars 100 forks source link

demo does not work. #79

Open mauriciovander opened 5 years ago

mauriciovander commented 5 years ago

I'm trying to run the demo in a Docker container. No luck so far

WARNING ([5.5.257~1416-73720]:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory.

These are the steps to reproduce the error:

Create a Dockerfile

FROM ubuntu

RUN apt-get update
RUN apt-get install -y git

RUN git clone https://github.com/kaldi-asr/kaldi.git kaldi --origin upstream

RUN apt-get install -y apt-utils g++ zlib1g-dev automake autoconf unzip wget sox libtool subversion python2.7 python3 libatlas3-base cmake libff$

RUN cd kaldi/tools; make -j `grep -c ^processor /proc/cpuinfo`

RUN cd kaldi/src; ./configure --shared

RUN cd kaldi/src; make depend -j `grep -c ^processor /proc/cpuinfo`
RUN cd kaldi/src; make -j `grep -c ^processor /proc/cpuinfo`

RUN git clone https://github.com/alumae/gst-kaldi-nnet2-online.git

RUN cd /gst-kaldi-nnet2-online/src; KALDI_ROOT=/kaldi make depend -j `grep -c ^processor /proc/cpuinfo`
RUN cd /gst-kaldi-nnet2-online/src; KALDI_ROOT=/kaldi make -j `grep -c ^processor /proc/cpuinfo`

Build an image and run the container

docker build -t ubuntu-kaldi-gstreamer .
docker run -d --name kaldi -p 80:80 ubuntu-kaldi-gstreamer
docker exec -it kaldi bash

Run the demo in the running container:

cd gst-kaldi-nnet2-online/demo/
export GST_PLUGIN_PATH=../src
./prepare-models.sh 
./transcribe-audio.sh dr_strangelove.mp3 

Don't get the expected JSON output, instead:

WARNING ([5.5.257~1416-73720]:LatticeWordAligner():word-align-lattice.cc:263) [Lattice has input epsilons and/or is not input-deterministic (in Mohri sense)]-- i.e. lattice is not deterministic. Word-alignment may be slow and-or blow up in memory.