Releases: jonasrauber/eagerpy
Releases · jonasrauber/eagerpy
Version 0.22.0
New Features
- reshape now accepts a single
int
(in particular-1
), not just tuples
Version 0.21.0
New Features
Tensor.__getitem__
now supports NumPy arrays (ndarray
)
Version 0.20.2
Bug fixes
- fixed a bug in
TensorFlowTensor.__getitem__
withslices
Version 0.20.1
New Features
- bug fixes for Python 3.7
Version 0.20.0
New Features
- added
Tensor.norms.*
methods- e.g.
x.norms.l2()
wherex
is anep.Tensor
instance
- e.g.
- other improvements to code, docs, tests
Version 0.19.0
New Features
- fully specified all parameters with descriptive names (no more
*args, **kwargs
) - fully type-annotated all functions and all tests
- improved docs
- bugfixes
Version 0.18.0
New Features
- documentation
- performance optimizations
__getitem__
now supportsrange
andlist
- added
take_along_axis
(for the last axis only) - better type annotations
Version 0.17.0
New Features
- more type annotations
ep.Tensor
now points to the base classep.TensorType
now points aTypeVar
bounded byep.Tensor
- added
astensors
,astensor_
andastensors_
- the underscored variants return a
restore_type
function that inverts theastensor
operation
- the underscored variants return a
- more tests
- stricter mypy config
Version 0.16.0
New Features
Tensor.tensor
is now calledTensor.raw
- added type hints to basically all methods and framework functions
- refactored large parts of the code
- many other bug fixes and improvements following from that
Version 0.15.0
New Features
TensorFlowTensor
now supports automatic casting to the commondtype
for all comparison operators- other improvements and bug fixes