tensorflow_hub library can be installed alongside TensorFlow 1 and
TensorFlow 2. We recommend that new users start with TensorFlow 2 right away,
and current users upgrade to it.
Use with TensorFlow 2
$ pip install "tensorflow>=2.0.0" $ pip install --upgrade tensorflow-hub
The TF1-style API of TensorFlow Hub works with the v1 compatibility mode of TensorFlow 2.
Legacy use with TensorFlow 1
tensorflow_hub library requires TensorFlow version 1.7 or greater.
We strongly recommend to install it with TensorFlow 1.15, which defaults to TF1-compatible behavior but contains many TF2 features under the hood to allow some use of TensorFlow Hub's TF2-style APIs.
$ pip install "tensorflow>=1.15,<2.0" $ pip install --upgrade tensorflow-hub
Use of pre-release versions
The pip packages
tf-hub-nightly are built automatically from
the source code on github, with no release testing. This lets developers try out
the latest code without building from source.
Optional: Setting the cache location for downloads.
tensorflow_hub uses a system-wide, temporary directory to cache
downloaded and uncompressed models. See Caching for options to use
other, possibly more persistent locations.
Although we hope to prevent breaking changes, this project is still under active development and is not yet guaranteed to have a stable API or model format.
As in all of machine learning, fairness is an important consideration. Many pre-trained models are trained on large datasets. When reusing any model, it’s important to be mindful of what data the model was trained on (and whether there are any existing biases there), and how these might impact your use of it.
Since they contain arbitrary TensorFlow graphs, models can be thought of as programs. Using TensorFlow Securely describes the security implications of referencing a model from an untrusted source.