core happened using chain model

[AjianIronSide]

I was trying to repeat the 'Practical Example' with a nnet3 chain model(tdnn_7q exactly) instead. I use the yaml file from sample_chinese_nnet3.yaml(multi_cn). However, a core file generated when decoder dealing with the audio chunk. There was no clue from worker or master log. Did I miss anything ?

[AjianIronSide]

Seems the same problem when using multi_cn model and yaml

[AjianIronSide]

The worker log is like:
2020-03-11 07:26:14 - INFO: decoder2: 67a2f518-a6ca-425b-8626-325027fc1599: Initializing request
2020-03-11 07:26:14 - INFO: decoder2: 67a2f518-a6ca-425b-8626-325027fc1599: Setting caps to audio/x-raw, layout=(string)interleaved, rate=(int)16000, format=(string)S16LE, channels=(int)1
2020-03-11 07:26:14 - INFO: decoder2: 67a2f518-a6ca-425b-8626-325027fc1599: Connecting audio decoder
2020-03-11 07:26:14 - INFO: main: 67a2f518-a6ca-425b-8626-325027fc1599: Started timeout guard
2020-03-11 07:26:14 - DEBUG: main: 67a2f518-a6ca-425b-8626-325027fc1599: Checking that decoder hasn't been silent for more than 10 seconds
2020-03-11 07:26:14 - INFO: main: 67a2f518-a6ca-425b-8626-325027fc1599: Initialized request
2020-03-11 07:26:14 - INFO: decoder2: 67a2f518-a6ca-425b-8626-325027fc1599: Connected audio decoder
2020-03-11 07:26:14 - DEBUG: main: 67a2f518-a6ca-425b-8626-325027fc1599: Got message from server of type <class 'ws4py.messaging.BinaryMessage'>
2020-03-11 07:26:14 - DEBUG: decoder2: 67a2f518-a6ca-425b-8626-325027fc1599: Pushing buffer of size 5120 to pipeline
2020-03-11 07:26:14 - DEBUG: decoder2: 67a2f518-a6ca-425b-8626-325027fc1599: Pushing buffer done

[hezhenke]

segment fault the same when use chinese test model