Trouble getting worker started about docker-kaldi-gstreamer-server

[jcsilva]

Hi,

I would like to ask you two things:

1. Please, show me all your worker.log and show me your yaml file.

2. Could you please try to run this example: https://github.com/jcsilva/docker-kaldi-gstreamer-server/tree/master/examples/practical-example. It should run without problems

[connorzwick]

Thank you for the quick response and apologies for my delay getting back to you!

I tried both of these yaml files:

nnet2.yaml.txt
modified.yaml.txt

And tail worker.log outputs:

python(PyEval_EvalFrameEx+0xae2) [0x4ca592]
python(PyEval_EvalCodeEx+0x411) [0x4c87a1]
python() [0x5030ef]
python(PyRun_FileExFlags+0x82) [0x4f8c72]
python(PyRun_SimpleFileExFlags+0x197) [0x4f7d77]
python(Py_Main+0x562) [0x4982f2]
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf5) [0x7fc9f2c47b45]
python() [0x497ca0]

And yes, the practical example ran without issue. Thanks!

[jcsilva]

The nnet2.yaml.txt will not work, because the paths "test/models" don't exist in the docker image.

Your modified.yaml.txt seems to be ok. Now, you should look the entire worker.log. Not just its tail. It is possible that you did not change the paths in the conf files (for example: all the paths in /opt/models/english/tedlium_nnet_ms_sp_online/conf/ivector_extractor.conf). In this case, you should see an [ERROR] message somewhere in worker.log.

[connorzwick]

You were right about changing the conf paths. I changed all of them and there is no longer the error message in worker.log. But now I am getting this:

https://gist.github.com/connorzwick/99203b4b9fc0a41b6225134e12cc8ed7

And the worker is still not live. Any ideas? Thanks for the help!

[wanghaifeng]

@connorzwick, I got same issue but now resolved by increase docker memory to 4G. You should try.

[jcsilva]

Hi, 1. for the first problem, I may just comment full-post-processor. So, it's gonna be #full-post-processor. 2. for your second problem, it seems your path to final.mat is wrong. In your container, if your model is located in /opt/model, the path to final.mat must start with /opt/model/... 2017-03-22 3:44 GMT-03:00 Prokopev Aleksey <[email protected]>:

…

Hi, first of all I also wanna thank @jcsilva <https://github.com/jcsilva> for creating this. It's awesome! But I also encountered problem with the worker. I downloaded the model and mounter "~/english" directory to "/opt/model/". Also I configured .yaml file to use files from this directory. I got rid of "libudev: udev_has_devtmpfs: name_to_handle_at on /dev: Operation not permitted " error by running docker with "--security-opt seccomp=unconfined". But it's kinda dirty solution. Here are log files and my .yaml configurations. master.log.txt <https://github.com/jcsilva/docker-kaldi-gstreamer-server/files/860548/master.log.txt> worker.log.txt <https://github.com/jcsilva/docker-kaldi-gstreamer-server/files/860547/worker.log.txt> nnet2.yaml.txt <https://github.com/jcsilva/docker-kaldi-gstreamer-server/files/860546/nnet2.yaml.txt> It seems like master got no problems but worker cannot find "./sample_full_post_processor.py" and also want to open "test/models/english/tedlium_ nnet_ms_sp_online/ivector_extractor/final.mat" for some reason. I didn't find the way to configure it. Now I'm looking for the way to handle it but it would be nice to know how to use it "off-the-shelf" later. Thank you a lot! — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#9 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFRWbHX2uRdmI0MIi7Zm_X9f0Mngs4kTks5roMM3gaJpZM4KXfNn> .