If the feature is related to a specific library below, please raise an issue in
the respective repo directly:
TensorFlow Data Validation Repo
TensorFlow Model Analysis Repo
TensorFlow Transform Repo
TensorFlow Serving Repo
System information
- TFX Version (you are using): 1.0.0-rc0
- Environment in which you plan to use the feature (e.g., Local
(Linux/MacOS/Windows), Interactive Notebook, Google Cloud, etc..): MacOS, AWS
- Are you willing to contribute it (Yes/No): Yes
Describe the feature and the current behavior/state.
tfx (1.0.0-rc0) currently depends on tensorflow-hub (>=0.9.0,<0.10)
I was wondering if we could update tensorflow-hub dependancy for tfx to allow tf-hub 0.12.0, so something like (>=0.9.0,<=0.12.0)?
I am not sure if that would break anything in tfx, but I am happy to investigate and contribute to this change
Will this change the current API? How?
No
Who will benefit with this feature?
tensorflow-hub has added some new features in 0.10.0 and beyond (specifically the one I'm interested in "compute_output_shape in hub.KerasLayer" which they added in 0.12.0). It would be cool to be able to take advantage of those while still being able to use tfx
Do you have a workaround or are completely blocked by this? :
Blocked
Name of your Organization (Optional)
Any Other info.
If the feature is related to a specific library below, please raise an issue in
the respective repo directly:
TensorFlow Data Validation Repo
TensorFlow Model Analysis Repo
TensorFlow Transform Repo
TensorFlow Serving Repo
System information
(Linux/MacOS/Windows), Interactive Notebook, Google Cloud, etc..): MacOS, AWS
Describe the feature and the current behavior/state.
tfx (1.0.0-rc0) currently depends on tensorflow-hub (>=0.9.0,<0.10)
I was wondering if we could update tensorflow-hub dependancy for tfx to allow tf-hub 0.12.0, so something like (>=0.9.0,<=0.12.0)?
I am not sure if that would break anything in tfx, but I am happy to investigate and contribute to this change
Will this change the current API? How?
No
Who will benefit with this feature?
tensorflow-hub has added some new features in 0.10.0 and beyond (specifically the one I'm interested in "
compute_output_shapeinhub.KerasLayer" which they added in 0.12.0). It would be cool to be able to take advantage of those while still being able to use tfxDo you have a workaround or are completely blocked by this? :
Blocked
Name of your Organization (Optional)
Any Other info.