tensor example support

[KwansuKim]

It seems that it is not possible to use "github.com/galeone/tfgo/proto/example" from tfgo v2.9. Are there any available ways to use example.Features and example.Example in tfgo v2.9? Or could you let me know another way to feed saved models without them?

[galeone]

Isn't possible to use tf.NewTensor and feed it as input to the saved model? I do this in the example you can see in the README.

Is this the thing you are looking for or do I misunderstand your question?

[galeone]

Now that I'm thinking about the proto/example thing clearly, I remember that I explicitly removed (a long time ago!) that part, because it's tensorflow itself that only supports SavedModels and the proto/example was there for feeding the estimators/frozen models. That type of model is not supported anymore by TensorFlow and therefore I removed them in this commit https://github.com/galeone/tfgo/commit/132f4f1c5dd7c604814ef7261fe40d3510e01220

You must use a SavedModel and feed the tensors using tf.NewTensor as shown in the readme

[KwansuKim]

Thanks for quick answer! I'm gonna try it according to your comment.

[KwansuKim]

@galeone Could you help me to resolve the below build issue? go: finding module for package github.com/tensorflow/tensorflow/tensorflow/go/core/framework/types_go_proto go: finding module for package github.com/tensorflow/tensorflow/tensorflow/go/core/protobuf/for_core_protos_go_proto go: finding module for package github.com/tensorflow/tensorflow/tensorflow/go/core/framework/tensor_shape_go_proto serving/ml/tensorflow imports github.com/galeone/tfgo imports github.com/tensorflow/tensorflow/tensorflow/go imports github.com/tensorflow/tensorflow/tensorflow/go/core/protobuf/for_core_protos_go_proto: module github.com/tensorflow/tensorflow@latest found (v2.9.1+incompatible), but does not contain package github.com/tensorflow/tensorflow/tensorflow/go/core/protobuf/for_core_protos_go_proto serving/ml/tensorflow imports github.com/galeone/tfgo imports github.com/tensorflow/tensorflow/tensorflow/go tested by github.com/tensorflow/tensorflow/tensorflow/go.test imports github.com/tensorflow/tensorflow/tensorflow/go/core/framework/tensor_shape_go_proto: module github.com/tensorflow/tensorflow@latest found (v2.9.1+incompatible), but does not contain package github.com/tensorflow/tensorflow/tensorflow/go/core/framework/tensor_shape_go_proto serving/ml/tensorflow imports github.com/galeone/tfgo imports github.com/tensorflow/tensorflow/tensorflow/go tested by github.com/tensorflow/tensorflow/tensorflow/go.test imports github.com/tensorflow/tensorflow/tensorflow/go/core/framework/types_go_proto: module github.com/tensorflow/tensorflow@latest found (v2.9.1+incompatible), but does not contain package github.com/tensorflow/tensorflow/tensorflow/go/core/framework/types_go_proto

[galeone]

You're using tensorflow, the official package - it's broken. You must use the fork I maintain instead:

github.com/galeone/tensorflow/tensorflow/go

As you can see the tf package in the example, comes from this package, not from the official one

[KwansuKim]

@galeone How can I make a tensor to input a model in case a model's input values involve different types like string, float32 at the same time? The following feature list is a sample I'd like to implement. With tensor.example, it would be possible to do it but now I'm not sure how to do it without tensor.example. "feature_a": "string_1" "feature_b": "string_2" "feature_c": 34.23 "feature_d": [1.0, 2.0, 3.0, 4.0]