Size of the model after training

[fly2mun]

I tried the train as Shmyrev said. It was difficult to set up the environment of kaldi but it was meaningful to me.

I Used this script https://github.com/alphacep/vosk-api/blob/master/android/lib/build-vosk.sh#L112

When the learning is over, the result binary is larger than the vosk website.

Training results are as follows. /exp/chain/tdnn/graph HCLG.fst size is 151,552KB phones.txt 2KB words.txt 2,856KB

If the script you shared is correct, I think a similar size should come out. I'm curious about your opinion.

Thanks, Moon.

[nshmyrev]

You probably used some other script, not the one to build kaldi

[fly2mun]

sorry.

https://github.com/alphacep/vosk-api/tree/master/training

and run.sh

[nshmyrev]

Ok, hclg size is expected then

[fly2mun]

Thank you so much. How was the content below created? Any scripts related to this?

This model seems to be small in size. vosk-model-small-en-us-0.15

Thanks, Moon.

[nshmyrev]

There is mkgraph_lookahead.sh script to make a dynamic graph.