[protobuf] tensorflow native data

[Philipp Kraus]

Hello, there,

I am currently working with Tensorflow for the first time and as far as I
know Tensorflow uses Protocol Buffers to store the data. I work on IoT
devices in a protected network and I have to deliver the data of a
Tensorflow DNNClassifier to external systems. These may be databases, but I
may also need to communicate with industrial controls as a byte stream.
Especially in this case I can only transfer primitive data like Int, Float,
Char etc. If I have now trained a Tensorflow DNNClassifier, a local
directory with the stored data will be created for me and I would now like
to read and process this data in native Python data types such as a Dict or
similar. How can I do this?
I also get a model as a native byte stream and would have to use it again
to generate a DNNClassifier. How can I realize this?

I am currently quite "new" with Tensorflow and Protocol Buffers.

Thank you very much for your help.

Phil

['Adam Cozzette' via Protocol Buffers]

You might want to reach out to the TensorFlow folks with your question. I
haven't really worked with TensorFlow much myself but it sounds to me like
what you're asking is more about TensorFlow than protobuf.

On Mon, Dec 3, 2018 at 2:54 PM Philipp Kraus <

Post by Philipp Kraus
Hello, there,
I am currently working with Tensorflow for the first time and as far as I
know Tensorflow uses Protocol Buffers to store the data. I work on IoT
devices in a protected network and I have to deliver the data of a
Tensorflow DNNClassifier to external systems. These may be databases, but I
may also need to communicate with industrial controls as a byte stream.
Especially in this case I can only transfer primitive data like Int, Float,
Char etc. If I have now trained a Tensorflow DNNClassifier, a local
directory with the stored data will be created for me and I would now like
to read and process this data in native Python data types such as a Dict or
similar. How can I do this?
I also get a model as a native byte stream and would have to use it again
to generate a DNNClassifier. How can I realize this?
I am currently quite "new" with Tensorflow and Protocol Buffers.
Thank you very much for your help.
Phil
--
You received this message because you are subscribed to the Google Groups
"Protocol Buffers" group.
To unsubscribe from this group and stop receiving emails from it, send an
Visit this group at https://groups.google.com/group/protobuf.
For more options, visit https://groups.google.com/d/optout.

[Philipp Kraus]

Hello, there,

I already asked this question in the Tensorflow group and there I was
referred to Protocol Buffer. But my question is still unanswered there.
Tensorflow itself already generates the appropriate data for me, I just
have to be able to read it out manually.
See the discussion at
https://groups.google.com/a/tensorflow.org/forum/#!topic/discuss/YMlRKvaEAoU

Thanks a lot

Phil

Post by 'Adam Cozzette' via Protocol Buffers
You might want to reach out to the TensorFlow folks with your question. I
haven't really worked with TensorFlow much myself but it sounds to me like
what you're asking is more about TensorFlow than protobuf.

Post by Philipp Kraus
Hello, there,
I am currently working with Tensorflow for the first time and as far as I
know Tensorflow uses Protocol Buffers to store the data. I work on IoT
devices in a protected network and I have to deliver the data of a
Tensorflow DNNClassifier to external systems. These may be databases, but I
may also need to communicate with industrial controls as a byte stream.
Especially in this case I can only transfer primitive data like Int, Float,
Char etc. If I have now trained a Tensorflow DNNClassifier, a local
directory with the stored data will be created for me and I would now like
to read and process this data in native Python data types such as a Dict or
similar. How can I do this?
I also get a model as a native byte stream and would have to use it again
to generate a DNNClassifier. How can I realize this?
I am currently quite "new" with Tensorflow and Protocol Buffers.
Thank you very much for your help.
Phil
--
You received this message because you are subscribed to the Google Groups
"Protocol Buffers" group.
To unsubscribe from this group and stop receiving emails from it, send an
<javascript:>.
Visit this group at https://groups.google.com/group/protobuf.
For more options, visit https://groups.google.com/d/optout.