Issue with building release version 0.1.1 DeepSpeech and Tensorflow

So I followed the steps, https://github.com/mozilla/DeepSpeech/blob/v0.1.1/native_client/README.md, to build tensorflow and deepspeech from source, from github.com/mozilla/Deepspeech and github.com/mozilla/tensorflow. I’m now trying to train my model but when I run DeepSpeech.py I get an error such that tensorflow module is not found. Any suggestions?

You should follow those steps: https://github.com/mozilla/DeepSpeech/tree/v0.1.1#training :slight_smile:

1 Like

Thanks but I did this instead since i’m building my own stuff

https://www.tensorflow.org/install/install_sources

Even if you are building your own stuff, you don’t have to build tensorflow python package. And if you build it yourself, then you need to share more information about the error, because “tensorflow module is not found” is not enough, but it’s likely this is not an error on our side.

Its okay. I solved it by installing the tensorflow bazel build as a pip package.