Changelog History
Page 2
-
v0.12.0-p0 Changes
February 26, 2020Tensor shapes and positions now explicitly track the tensor's rank.
๐ฅ breaking changes:
fdeep::tensor5
has been renamed tofdeep::tensor
fdeep::tensor5_pos
has been renamed tofdeep::tensor_pos
fdeep::shape5
has been renamed tofdeep::tensor_shape
- โฌ๏ธ dropped support for shape inference in
reshape
layers
๐ deprecated functions (will likely be removed from the API soon)
float_type fdeep::tensor5::get(std::size_t, std::size_t, std::size_t, std::size_t, std::size_t) const
: Please usefloat_type fdeep::tensor5::get(const tensor_pos&) const
orfloat_type fdeep::tensor5::get_ignore_rank(const tensor_pos&) const
instead.void fdeep::tensor5::set(std::size_t, std::size_t, std::size_t, std::size_t, std::size_t, float_type)
: Please usefloat_type fdeep::tensor5::set(const tensor_pos, float_type)
orfloat_type fdeep::tensor5::set_ignore_rank(const tensor_pos&, float_type)
instead.
-
v0.11.1-p0 Changes
December 17, 2019- ๐ Support for batch normalization on arbitrary axes
- ๐ Improved error messages
-
v0.11.0-p0 Changes
November 07, 2019Switch from
keras
totf.keras.
๐ With the release of version 2.3.0, team Keras announced the following:
๐ > This is also the last major release of multi-backend Keras. Going forward, we recommend that users consider switching their Keras code to tf.keras in TensorFlow 2.0. It implements the same Keras 2.3.0 API (so switching should be as easy as changing the Keras import statements), but it has many advantages for TensorFlow users, such as support for eager execution, distribution, TPU training, and generally far better integration between low-level TensorFlow and high-level concepts like Layer and Model. It is also better maintained.
๐ > Development will focus on tf.keras going forward. We will keep maintaining multi-backend Keras over the next 6 months, but we will only be merging bug fixes. API changes will not be ported.
So frugally-deep follows this direction.
Thanks to @keithchugg for doing the needed adjustments of the stateful implementations.
-
v0.10.1-p0 Changes
October 30, 2019๐ Improved tests and docs.
-
v0.10.0-p0 Changes
October 29, 2019- ๐ Support for stateful models, supporting GRU and LSTM layers. ๐
Huge thanks to @keithchugg for the amazing work he invested in this. ๐ฅ
-
v0.9.9-p0 Changes
October 29, 2019- ๐ Fix edge-case behavior in softmax implementation
-
v0.9.8-p0 Changes
October 06, 2019๐ Fix behavior or
Upsampling1D
for input tensors with different ranks. -
v0.9.7-p0 Changes
September 21, 2019- ๐ Improve handling of sequential models
- ๐ Better documentation for image conversion
- New convenience function
model::predict_class_with_confidence
-
v0.9.6-p0 Changes
September 09, 2019- ๐ Fix concatenation of tensors with
axis < rank
- ๐ Fix concatenation of tensors with
-
v0.9.5-p0 Changes
September 09, 2019- ๐ Fix linker issues