Sathya R
Facing similar error https://github.com/tensorflow/tensorflow/pull/12759#issuecomment-467860111
> @davideboschetto No, you don't. I found [this](https://stackoverflow.com/questions/35508866/tensorflow-different-ways-to-export-and-run-graph-in-c/43639305#43639305) discussion very helpful. You can load checkpoint file to C++ and do inference thereafter. @davideboschetto @hjchai 1. Loading Graph/Checkpoint files in C++...
It seems beam search functionality could not be added to inference graph incase if you need to use in C++ or TF Serving. Kindly refer this [issue](https://github.com/google/seq2seq/issues/354)
If so, could some one point out which function exactly should I convert to TF Ops (the above error message could possibly help)? `py_func` is used [here](https://github.com/google/seq2seq/blob/7f485894d412e8d81ce0e07977831865e44309ce/seq2seq/inference/beam_search.py#L90)
I have been trying to implement this function in pure TF ops, ``` def gather_tree_py(values, parents): beam_length = values.shape[0] num_beams = values.shape[1] res = np.zeros_like(values) res[-1, :] = values[-1, :]...
> you could just comment out the related code: line 68~73 of attention_seq2seq.py and "if self.use_beam_search: **" of basic_seq2seq.py to avoid using beam search since the related code almost based...
is that possible to use this method defined in [tensorflow](https://www.tensorflow.org/guide/saved_model#using_savedmodel_with_estimators) with seq2seq? Or is there built-in support to **export model** by seq2seq? There is also something like [_maybe_export](https://github.com/tensorflow/tensorflow/blob/r1.12/tensorflow/contrib/learn/python/learn/experiment.py) in official...
> Hi guys, I come across the same problem and try to solve this in another way. > When training within MonitoredTrainingSession, I use a tf.train.SessionRunHook to save the model...
@amirj @gidim do you think the approach followed [here](https://stackoverflow.com/questions/52120188/how-to-get-tensorflow-served-model-to-pull-from-passed-in-input-and-not-local-ba) to serve the graph is recommended?
> Should figure out how to export models for serving, I think Tensorflow does provide something like an `ExportStrategy` that can be passed to the estimator and it will occasionally...