Tricks (deprecated)
Warning
The tricks
module is deprecated since OTBTF 2.0
The Tensorflow python API has changed significantly (for the best) after the
Tensorflow 2.0 release. OTBTF used to provide the tricks
module, providing
useful methods to generate the SavedModels, or convert checkpoints into
SavedModels.
What is best in Tensorflow 2.X?¶
- Shorter and simpler code
- Easy to build and train a model with Keras, which has become the principal interface for Tensorflow
- More hardware-agnostic than ever: with the exact same code, you can run on a single-cpu, GPU, or a pool of GPU servers.
Major changes between Tensorflow 1 and Tensorflow 2 APIs¶
Here are a few tips and tricks for people that want to move from Tensorflow 1 to Tensorflow 2 API. Models built for OTBTF have to take in account the following changes:
- Models built with
otbtf.ModelBase
ortensorflow.keras.model.Model
have no longer to usetensorflow.compat.v1.placeholder
buttensorflow.keras.Input
instead, - Tensorflow variable scopes are no longer used when the training is done from Keras,
- SavedModel can be created directly from the model instance! (as simple as
mymodel.save("mymodel_v1_savedmode")
), - Switching between single cpu/gpu or multiple computing nodes, distributed
training, etc. is done using the so-called
tensorflow.Strategy
Note
Read our tutorial to know more on working with Keras!