How do I export an eager execution model?
up vote
1
down vote
favorite
Having completed my model, I now wish to export and deploy it, following this tutorial on TensorFlow's website. However, there is no indication of how to do this in eager execution, where I am unable to provide a session or graph to builder.add_meta_graph_and_variables().
Is this a case where my code needs to be eager and graph compatible, or where I need to save my model, import it to a session, and export it from there?
tensorflow tensorflow-serving
asked Nov 9 at 17:04
Jordan Patterson
816
add a comment |
up vote
1
down vote
favorite
Having completed my model, I now wish to export and deploy it, following this tutorial on TensorFlow's website. However, there is no indication of how to do this in eager execution, where I am unable to provide a session or graph to builder.add_meta_graph_and_variables().
Is this a case where my code needs to be eager and graph compatible, or where I need to save my model, import it to a session, and export it from there?
tensorflow tensorflow-serving
asked Nov 9 at 17:04
Jordan Patterson
816
1
See Work with graphs in the Eager Execution guide. As it says there, "deploying code written for eager execution is more difficult", and you will need to do some extra work for yourself. That said, eager mode will become even more relevant in the future, and we may see improvements in that area.
– jdehesa
Nov 9 at 17:15
Thanks for the reply jdehesa! I am looking forward to 2.0, but it probably isn't releasing till December. Between "either generate a graph from the model, or run the Python runtime and code directly on the server," I guess I will try for the former. This seems like its the same problem I have raised with this, as many methods only work for either graph or eager, not both.
– Jordan Patterson
Nov 9 at 17:32
add a comment |
up vote
1
down vote
favorite
up vote
1
down vote
favorite
Having completed my model, I now wish to export and deploy it, following this tutorial on TensorFlow's website. However, there is no indication of how to do this in eager execution, where I am unable to provide a session or graph to builder.add_meta_graph_and_variables().
Is this a case where my code needs to be eager and graph compatible, or where I need to save my model, import it to a session, and export it from there?
tensorflow tensorflow-serving
asked Nov 9 at 17:04
Jordan Patterson
816
Having completed my model, I now wish to export and deploy it, following this tutorial on TensorFlow's website. However, there is no indication of how to do this in eager execution, where I am unable to provide a session or graph to builder.add_meta_graph_and_variables().
Is this a case where my code needs to be eager and graph compatible, or where I need to save my model, import it to a session, and export it from there?
tensorflow tensorflow-serving
tensorflow tensorflow-serving
asked Nov 9 at 17:04
Jordan Patterson
816
asked Nov 9 at 17:04
Jordan Patterson
816
asked Nov 9 at 17:04
Jordan Patterson
816
asked Nov 9 at 17:04
Jordan Patterson
816
asked Nov 9 at 17:04
Jordan Patterson
816
816
1
See Work with graphs in the Eager Execution guide. As it says there, "deploying code written for eager execution is more difficult", and you will need to do some extra work for yourself. That said, eager mode will become even more relevant in the future, and we may see improvements in that area.
– jdehesa
Nov 9 at 17:15
Thanks for the reply jdehesa! I am looking forward to 2.0, but it probably isn't releasing till December. Between "either generate a graph from the model, or run the Python runtime and code directly on the server," I guess I will try for the former. This seems like its the same problem I have raised with this, as many methods only work for either graph or eager, not both.
– Jordan Patterson
Nov 9 at 17:32
add a comment |
1
See Work with graphs in the Eager Execution guide. As it says there, "deploying code written for eager execution is more difficult", and you will need to do some extra work for yourself. That said, eager mode will become even more relevant in the future, and we may see improvements in that area.
– jdehesa
Nov 9 at 17:15
Thanks for the reply jdehesa! I am looking forward to 2.0, but it probably isn't releasing till December. Between "either generate a graph from the model, or run the Python runtime and code directly on the server," I guess I will try for the former. This seems like its the same problem I have raised with this, as many methods only work for either graph or eager, not both.
– Jordan Patterson
Nov 9 at 17:32
1
1
See Work with graphs in the Eager Execution guide. As it says there, "deploying code written for eager execution is more difficult", and you will need to do some extra work for yourself. That said, eager mode will become even more relevant in the future, and we may see improvements in that area.
– jdehesa
Nov 9 at 17:15
See Work with graphs in the Eager Execution guide. As it says there, "deploying code written for eager execution is more difficult", and you will need to do some extra work for yourself. That said, eager mode will become even more relevant in the future, and we may see improvements in that area.
– jdehesa
Nov 9 at 17:15
Thanks for the reply jdehesa! I am looking forward to 2.0, but it probably isn't releasing till December. Between "either generate a graph from the model, or run the Python runtime and code directly on the server," I guess I will try for the former. This seems like its the same problem I have raised with this, as many methods only work for either graph or eager, not both.
– Jordan Patterson
Nov 9 at 17:32
Thanks for the reply jdehesa! I am looking forward to 2.0, but it probably isn't releasing till December. Between "either generate a graph from the model, or run the Python runtime and code directly on the server," I guess I will try for the former. This seems like its the same problem I have raised with this, as many methods only work for either graph or eager, not both.
– Jordan Patterson
Nov 9 at 17:32
add a comment |
active
oldest
votes
active
oldest
votes
active
oldest
votes
active
oldest
votes
active
oldest
votes
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f53230290%2fhow-do-i-export-an-eager-execution-model%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
1
See Work with graphs in the Eager Execution guide. As it says there, "deploying code written for eager execution is more difficult", and you will need to do some extra work for yourself. That said, eager mode will become even more relevant in the future, and we may see improvements in that area.
– jdehesa
Nov 9 at 17:15
Thanks for the reply jdehesa! I am looking forward to 2.0, but it probably isn't releasing till December. Between "either generate a graph from the model, or run the Python runtime and code directly on the server," I guess I will try for the former. This seems like its the same problem I have raised with this, as many methods only work for either graph or eager, not both.
– Jordan Patterson
Nov 9 at 17:32